• 5 hours ago
Will AI replace data analysts—or change what the role looks like? Alex Freeberg explores the state of AI in analytics, weighing the risks of automation against the skills and human judgment businesses still need.

The discussion looks at ChatGPT’s Code Interpreter, Microsoft Copilot, and AI agents, including their potential to speed up data analysis and their limitations with errors, messy data, and business context. Alex shares predictions about how AI may reshape analytics careers, create more specialized roles, and increase demand for people who can build, check, and maintain AI-powered workflows. This data analytics commentary is for aspiring and working analysts researching AI, automation, ChatGPT for data analysis, and the future of data careers.

Category

📚
Learning
Transcript
00:00:00Hello, everybody. I am your host, Alex Freeberg, and this is the Alex the Analyst Show. Thank you
00:00:03so much for joining me. Today, we're going to be talking about the state of AI in analytics.
00:00:09Now, there is a lot to cover, and I mean, I have my notes here. I have pages of notes
00:00:14on just about
00:00:15everything. We're going to be talking about things like ChatGBT's Code Interpreter,
00:00:20Microsoft's Copilot 365, which should be released in not too long. We're going to be talking about
00:00:24the positives and the negatives that I personally see with just artificial intelligence in general
00:00:30and analytics. Then we're going to be talking about some of my predictions and kind of where I
00:00:35think a lot of people should be focusing a lot of their time and energy in the coming months and
00:00:41years as these things progress. So there's a lot to talk about. Again, if you haven't seen one of
00:00:48these shows, now, this is not like my regular episodes where they're like 15 minutes or less.
00:00:51This is my Alex the Analyst Show, where I just talk, and we just have a conversation. You can
00:00:56grab some coffee. I just have water today. I usually have like tea or something, but we just
00:01:01chat, and this one to me is extremely important. Just to let you guys know, I had had something
00:01:08planned to talk about maybe two months ago. I had it all written out, and I kind of started
00:01:14second-guessing myself. I was like, I need to do more research. I need to really dive into this more
00:01:18because I don't want to come to you guys just, you know, with my initial knee-jerk reaction and
00:01:24just kind of tell you what I'm thinking. I really wanted to have, especially for something as
00:01:27important as this, I really wanted to have some, just a better perspective on things and really
00:01:33be grounded in what I believe and what I think in this area. It's been a very interesting past six
00:01:40months. I've seen, I really, I hear, I get a lot of feedback from the community, whether it's
00:01:47LinkedIn, Twitter, Instagram, YouTube. I get a lot of feedback, and there's just been a lot of worry,
00:01:54a lot of, you know, what's happening is like, are jobs okay? Should I still be studying? Should I still
00:02:00be learning these things, or should I just choose a totally different career path? That's my cat.
00:02:05Uh, and those are all very, very normal things to think about. Honestly, if I was at the very
00:02:13beginning of my career right now, um, and I didn't know as much as I know now, I might have
00:02:17a very
00:02:18different outlook on things. So I have a lot of information. My cat's being weird right now.
00:02:22She's like staring at me through the crack in my door, but I have a lot of information to cover
00:02:26a lot.
00:02:27So just stick with me, hang out. Um, I think it'll be a really interesting conversation. I'm going to post
00:02:32a
00:02:32lot of this stuff to LinkedIn as well. Um, you know, when this goes live, but, um, with that being
00:02:37said, let's kind of start off with just, you know, what are things looking at right now? Now, as of
00:02:44this recording, it's May 8th of 2023, I'll probably release this next week. Um, but you know, it's, um,
00:02:52it's May basically things have changed so much, even in previous months, two months ago, when I was
00:02:58writing this, some of these things weren't announced yet. Things like Microsoft copilot
00:03:03365, uh, things like chat GPT's code interpreter. These things are just recently coming out and
00:03:09there's other things like auto GPT, baby GPT, uh, and some other things that we're going to talk
00:03:14about as well. So really quickly, I want to mention some of these technologies, what they're doing,
00:03:19why people are concerned about it. Um, and then I'm going to kind of really go into like my positive
00:03:25negatives on these types of technologies and what I think is actually going to happen. Those
00:03:29are my predictions. First thing I want to talk about is, uh, chat GPT's code interpreter. Now,
00:03:33uh, usually I don't do any overlays. I'm leaning forward so much. Um, normally I don't do any
00:03:39overlays in these types of things. I just tell it how it is. I'm going to try to give you
00:03:42a little
00:03:43demonstration of what this is, uh, on the screen in front of you, but chat GPT's code interpreter is
00:03:48chat GPT's newest, uh, kind of plugin integration that they're releasing. And a lot of people are saying
00:03:53it's basically a data analyst for $20 a month. Um, and so it's, you know, it has a lot of
00:04:00enhanced abilities. You can, uh, feed it a lot of different things. Things like files are going to
00:04:06be able to start feeding into it. Um, and so that's, and they're saying it's a data scientist for $20
00:04:11a
00:04:11month. So they're basically like, you know, that's what it is. It's supposed to be somebody who analyzes
00:04:16data, which is really interesting. Then we have Microsoft copilot 365, which is basically just
00:04:21chat GPT inside of all of Microsoft's products. Uh, so it's going to be in, uh, let me see word,
00:04:29Excel, PowerPoint, teams, outlook, and more as you know, they integrate these things.
00:04:33They'll probably be in Azure as well. Power BI, all their technologies really. Then you have things
00:04:39like auto GPT and baby GPT. Now this one's a little bit more niche. Not as many people know about
00:04:44this
00:04:44one, but basically the premise of it is that you are creating these little bots with chat GPT,
00:04:51functionality. And these little agents is what they call them. And the agents go out and they
00:04:55do a specified task. And then when they're done, they are done, right? But you can have multiple
00:05:00of them and the agents themselves can create like sub agents to do smaller tasks. And so you can tell
00:05:06it, Hey, go and do this entire task and do this, this, and this, and this, and it'll go and
00:05:11do all
00:05:11of those things and it'll come back and it'll be done. So really, really, um, interesting. I've
00:05:16actually kind of dabbled in that myself. I got it onto my computer. I set everything up. Um, it was
00:05:22pretty tough. In my opinion, people were making it seem really easy online. Um, but I actually
00:05:27found it quite difficult with that being said, these are all technologies that are like just kind
00:05:32of starting to come out. Um, and these are the types of things that I think people are most worried
00:05:38about is some type of, you know, one product that can literally just be a data analyst. Um, and so
00:05:46I'm going to be talking about that in a little bit. Um, but that is, as of right now, that
00:05:50is like,
00:05:51that's where we're at. That's the state of AI and analytics is people are just worried. Um, and that's
00:05:57on the, you know, our side of things, which is the employees, the people who want to become data
00:06:01analysts, who are a data analyst, you know, will they have a job in six months, a year, five years,
00:06:06will there be jobs we're going into if you're getting a degree in data analytics or a degree
00:06:11in data science or these things, you know, are there going to be jobs for this in the future?
00:06:17Again, extremely, extremely, extremely understandable. Um, what I want to go into right
00:06:22now is kind of the negatives. Um, the, the things that I see are negative to a good outcome
00:06:28with these types of technologies for data analysts like us. I'm just going to preface this by saying I
00:06:34have a lot of positives. I have some really interesting predictions. Um, but the negatives
00:06:39I think are very real, right? I think they're pretty obvious too. And I think once I start talking
00:06:44about them, you're going to be like, if you haven't thought about them before, you're going to be like,
00:06:46that makes a lot of sense. Uh, the first thing is automation of repetitive tasks. That's what a lot
00:06:54of, um, chat GPT's ability to automate things. It's just these repetitive tasks, you know, think,
00:07:00um, think like an accountant or a lawyer or somebody who writes something you're writing,
00:07:06you're generating these things. And it's a lot of repetitive things, something that chat GPT is
00:07:11really, really good at. And so, um, you know, one of the biggest things that a lot of people pointed
00:07:17to or the negative side, and I can see this is that there actually is a lot of automation in
00:07:22analytics. I did it myself. When I first got into one of my previous roles, I was like, man, a
00:07:26lot of
00:07:27this can be automated. Uh, and that's what I started doing, uh, mostly with Python, but you
00:07:32know, what a lot of people will say, a lot of people are arguing, and some of it I agree
00:07:36with
00:07:36is that there are a lot of automatable, uh, if that's a word, there's a lot of things that can
00:07:41be automated in analytics, um, for the good or the bad. I mean, you know, automation is, is very,
00:07:49very, very useful, right? You don't have to do all these things manually. And so now you can
00:07:54automate something to happen and happens over and over again with consistency, right? That's
00:08:00a fantastic thing. But, you know, one of the concerns, um, is that data analysts is just
00:08:08so much repetitive work. So many things that can be automated that literally you won't need
00:08:12a data analyst anymore. Um, and I'll get to that. Uh, I'll come back on that in a little
00:08:17bit. The next thing, kind of the next negative, the next thing that I see is that, um, chat GPT
00:08:23can actually do pretty good data analytics. Now I am going to come about it from my own
00:08:30personal perspective, some negative and some positive. The positive is that it can do advanced
00:08:35data analytics. I've seen it. Um, I've tried it out. I've messed with it. Um, I have a whole
00:08:40video out on chat GPT for data analytics, and that's just like a basic, here's how you can
00:08:45kind of use it, but I've seen demonstrations. I've also kind of gone and dived into it myself
00:08:50to do some more complex things. And this is things like the code interpreter, chapter
00:08:55GPT's code interpreter. That's what it's really, really good for is doing these a lot more complex
00:09:01things. Um, where it is right now today, it's not perfect. Uh, it just isn't. And it can
00:09:09do a lot of things with basic information, um, which is good, right? It's, it's a good thing
00:09:16that it can do this basic information. Now in a year, three years, five years, it's probably
00:09:21going to get better. That's my worry is that it gets so good that, uh, at least that's like
00:09:27the, this kind of argument is that gets so good where it's like most people wouldn't need a data
00:09:31analyst. You'll just go to this code interpreter, plug in the data that you have, ask a few questions
00:09:35and you're done. Um, to kind of support that, cause I'm going to kind of play devil's advocate
00:09:42a little bit here to kind of support that some companies, some businesses, that's all they would
00:09:46really need. Um, they wouldn't need a full on, um, you know, a full on network of databases. They
00:09:54don't need a lot of these other things. They just need somebody to analyze some excels for them, or they
00:09:58need somebody to analyze some data sitting in a simple database. That's where I think this is
00:10:03actually going to come in really well for those people and might reduce the need for additional
00:10:08people to come in and analyze their data when they can do something like this. And their data is like
00:10:13not crazy, crazy complex. Um, or even if it is somewhat complex, like this stuff could potentially
00:10:18do that, uh, in the future. Now I messed around with it. Um, I watched a few videos on it
00:10:24and
00:10:25honestly it's impressive, but I noticed there were several errors. I was having to prompt it to fix
00:10:30errors. I was having to prompt it to understand what the data actually was within the columns.
00:10:35So it's not perfect, right? It just isn't, but it still is really impressive. This is something that
00:10:41I could see a lot of people using even more than just the standard chat GPT. Um, and, and especially
00:10:47as plugins come along and not something I didn't really talk a lot about, but plugins are, you know,
00:10:52allow chat GPT to plug in to these different applications. Eventually most of the things on the
00:10:58internet potentially, um, or even applications. Like I could see Snowflake and I could see, uh,
00:11:03Databricks and Azure and all these things have these plugins where you can sit chat GPT on top
00:11:07of these things and it can pull in that data or read that data in and do a lot of
00:11:11things for it.
00:11:12So, um, that is definitely a potential negative, a potential, uh, bad thing when it gets so advanced,
00:11:18it could potentially get so advanced where it does a lot of this work. Uh, the next thing is cost
00:11:23effectiveness and efficiency. It is no, uh, it's no surprise that data analytics and having a
00:11:30department on data, uh, analysts and employing data analysts and keeping up the software and everything
00:11:36is expensive, right? Most people are expecting around 60,000 at the lower, lower level entry level,
00:11:42um, 60,000 a year in us dollars for employment. Um, most people want more. They want 80,000,
00:11:49100,000 for senior levels. You're going to get 125,000. Then you have managers. Then you have
00:11:54all of your tech stacks and you have all of your cloud compute. You have, um, your software license
00:11:59that you need. Like there's a lot that goes into that and there's recruiting for those positions.
00:12:03There's a lot of overhead for data analytics, especially at large companies. There's just a
00:12:07lot of overhead. The biggest negative to something like this is that it's very cheap or negative to
00:12:14people like us, data analysts, people, data scientists, uh, people in the, that work in this data,
00:12:19uh, environment is that it's really, really cheap, uh, to actually run ChatGPT for like OpenAI or ChatGPT
00:12:29or a lot of these other companies. It's crazy cheap and it can do a lot of work very quickly.
00:12:34So it's extremely cost effective. It's extremely efficient for a lot of the overhead stuff. Let's
00:12:42say it can, let's say right now it can do 50% of what a data analyst could do. Let's
00:12:46just say that's
00:12:47essentially, you know, $30,000 a year and they might pay like a thousand dollars for it to do
00:12:53all of these things. So, I mean, that's just some random numbers, right? But if it's doing half of
00:12:57what you're needing to be doing and it's doing it extremely cheaply, um, that's a crazy thing for
00:13:04like employers to ignore, right? For companies to ignore. That's really serious. Um, so that's,
00:13:11uh, you know, that's just, I didn't talk about efficiency so much, but it can do a lot. Let's
00:13:17say like it can do it a thousand times faster. And let's say I can do an end to end
00:13:21project
00:13:21that you would do for, you know, it would take you two weeks and it could do it in 10
00:13:27minutes,
00:13:28right? That, that for example, would be a huge reason why someone would get this type of system
00:13:34in their company and they would start using it because they can get things done so much faster,
00:13:39incredibly fast. That's something that a human just can't do. Uh, so that is a huge positive
00:13:44for automation and a negative for data analysts. Next thing is continuous advancements in the
00:13:50technology, right? We're, we're very early on. Um, this will, I'll talk a lot about this in my
00:13:57predictions section later, but there's a lot of things that are going to happen in the next several
00:14:03years, just within the past six months. So much has happened. Um, it's, it's been really tough to
00:14:09keep up with even for myself. And I keep up with it a lot. Like I'm always on top of
00:14:14it. I'm, I'm,
00:14:15I'm leaning in a lot today. I'm like this. Um, there's a lot that's come out and these are just
00:14:21to name a few, but things are escalating at a pace. I, I really haven't seen before. Um,
00:14:26I think it's just because this technology is so, it's so, uh, uh, it's not unique to one area.
00:14:35It's being used in all areas. I think that's, what's so unique about it's, it's every type of
00:14:42work, whether you're in paralegal, whether you are in, you know, um, uh, accounting, banking,
00:14:50fintech, whatever, there's a, every, everybody's using it in some way. Um, and it's spreading
00:14:56so quickly. It's like, everybody is trying to create the next best thing. And again,
00:15:02I don't know how to describe it because it's just fascinating how incredibly fast this is moving.
00:15:07Um, and so, you know, as these things progress, they may get better and better and better and
00:15:12better and better and even better than they are today. I'm sure they will be in a year.
00:15:16There'll be a lot better than they are today. So that's a genuine worry. Um, because we can't
00:15:22keep up data analysts can not learn as fast as AI is going to learn ever. That's a very genuine
00:15:29concern. Um, so to kind of summarize, at least the negatives, uh, and I call them negatives,
00:15:37but they're negatives for us and they're positives for others, right? Small business owners, companies,
00:15:44they can save a ton of money. Uh, they can do a lot of things where you don't need the
00:15:48overhead of
00:15:49and, uh, hiring somebody to do it. It'll be a lot faster, a lot cheaper. Um, I mean,
00:15:54that's essentially what AI is promising, right? And that's what everyone's worried about. They're
00:15:58like, goodness gracious. I mean, you know, if this technology is, it does what people are hyping
00:16:05it up to do. Um, I'm not going to have a job. Uh, you know, so that's, it's a really
00:16:11tough pill to
00:16:12swallow, uh, to really, I'm, and I'm trying to be like as real as I possibly can be here. Um,
00:16:18just because this is a very serious topic. I'm smiling, I'm laughing, but this is an extremely
00:16:23serious topic, um, legitimately. And so now I kind of want to go away from the negatives for just a
00:16:28little bit. It'll kind of be in there, but I want to talk more about the positives of AI, um,
00:16:35just for job stability in general, um, as a data analyst, not as an employer as a data analyst.
00:16:44Here's some of the positives when I'm looking at AI, um, when I'm working with these technologies
00:16:49and I've been testing out a ton, I keep, I want to reiterate that. Like I, I'm not just
00:16:54seeing stuff online and kind of like spitting it back out. Like I'm testing all of these things
00:16:58myself. Here's one of the biggest positives, uh, for us is right now AI is not perfect.
00:17:05It's going to make a lot of mistakes. Um, and when I say a lot of mistakes on the surface
00:17:12level
00:17:12from everything that I've seen, everything that I've tried, um, not just try, but all the
00:17:17demonstrations that I've seen is it always makes some mistake. Um, on the very bottom level,
00:17:24like the lowest level, simple analytics, it gets it right. And it's not, it's not even, uh,
00:17:32it's not unimpressive. It's still impressive, but that type of technology has actually been
00:17:36around for a while. Um, and I have seen it out there and people even before chat GPT were worried
00:17:42about these types of technologies, you know, coming into the main, uh, area and taking over.
00:17:49And I'm like, I've seen these things. Um, and they've been around for, since I've been a day
00:17:54for the past six years, they've been around where they'll ingest some data, excuse me,
00:18:00they'll do some analytics. They'll spit out some visualizations that actually is not new. Uh,
00:18:05chat GPT is just making a lot more center stage. Right. And so when you watch something like, and I,
00:18:11I suggest you look into these things and I'll try to leave links in the description for all of these,
00:18:15um, like demos and stuff like that. But chat GPT's code interpreter is really impressive. Um,
00:18:21but when you get your hands on it, when you start using it, you'll notice it makes mistakes. Um,
00:18:27and that is a really, really strong positive for us, right? As data analysts, as data scientists,
00:18:33people in the data world is it's going to make mistakes. Um, and let's say in five years,
00:18:3810 years, it's making a lot less mistakes. It's still going to make mistakes. Um, and for a lot of
00:18:43companies, these little mistakes could be huge. Uh, you know, if you rely so much on this auto,
00:18:51this automation, these analytics, you're just kind of blindly trusting it. That's a problem
00:18:56because if it starts making any type of mistake and you start blindly trusting, it could lead to
00:19:00millions of dollars. I mean, literally millions of dollars in lost revenue in, in, um, not just
00:19:06revenue, but you know, we're talking about client retention. We're talking about saving money,
00:19:11spending money, how you allocate your resources. There's a lot of things that analytics, uh, goes
00:19:15into. So, you know, just the fact that it's making mistakes, even on not crazy complicated stuff,
00:19:22or doesn't fully understand it, or you really have to like dig in to make it understand or to make
00:19:27it
00:19:28know what you're trying to tell it, you know, it takes a lot of, uh, uh, it makes mistakes and
00:19:34takes
00:19:34effort. Um, the next thing that's, uh, kind of a positive, at least I definitely see this as a
00:19:39positive is that AI can't do everything. Uh, and it needs a lot of context and understanding
00:19:45from someone like me, uh, who's inputting it. It needs a lot of understanding to really understand
00:19:50it. Now, again, I'm pointing back to the code interpreter because that one's actually somewhat
00:19:54new. Um, but I was able to get my hands on it in the code interpreter. Um, it doesn't need
00:19:59a lot
00:19:59of context. It tries to extrapolate as much information as it can from the information and
00:20:04the data that you've given it. Now, if that data is well labeled, if it comes with documentation,
00:20:11um, and you give it a little bit of context, it does a really good job. The funny thing is
00:20:18though,
00:20:18is I have not really ever worked in an environment where that is the case. Um, and that's the positive,
00:20:26right? Um, the positive is just humans, humans are human and we make mistakes. Therefore AI is not
00:20:34going to like, just be really good at this stuff until it's AI everywhere in every department doing
00:20:42everything, which, um, I just, I, you know, I don't see that happening, uh, anytime soon. Right.
00:20:51So what I'm trying to get at is, is that humans are feeding chat GPT. They're feeding AI what it
00:20:58needs
00:20:58to know in order to perform these analytics and humans just get things wrong. There's dirty data and,
00:21:04and things like code interpreter can actually clean data, but if they don't understand the
00:21:09nuances of the data, they can't do it well. Uh, and it makes mistakes. That's, you know,
00:21:14just going back, it makes mistakes. And so, you know, it's really interesting that, you know,
00:21:21probably the biggest impediment, the biggest roadblock for AI is just humans itself. Um, we,
00:21:28we just make a lot of mistakes. Um, and you know, humans are always going to be part of the
00:21:35process.
00:21:35I worked with healthcare, right? Humans are what's inputting data that makes mistakes. Humans are the
00:21:42ones, you know, um, working on the front lines and putting in notes and, and putting in things and
00:21:49doing these codes and all these, uh, different types of things within healthcare. And they just
00:21:54make a lot of mistakes. And so you can train it on what to look for and automate certain processes,
00:22:01but there are always going to be other errors. There are always going to be different things
00:22:05that they need to figure out and learn from and train from. And then those analytics, there's a
00:22:09lot of nuances in business. Um, this actually may be my next point. Uh, it's actually not. So I'm
00:22:16going to talk about it is that every business or every use case within a business is different.
00:22:22So from what I've seen from a code interpreters, you can give it the data, you can ask it the
00:22:28questions and it'll give you some information. But when you start getting a lot more specific on
00:22:32certain information that you're wanting to get out of it, it sometimes is, it needs more information
00:22:40and you're trying to ask certain questions and then you're like, okay, I need more data.
00:22:45I need more data. I need data from here. I need data from there. So it can't give you everything.
00:22:49Just you have to feed it the right information. And with that being said, a business like the one
00:22:56I was working at, we had a hundred different clients, all wanting unique things. They all
00:23:00wanted different things to be done with their data. And so, uh, let me see how much time I got.
00:23:05I'm at 25 minutes. So I'm, or, or so, so I'm, I'm just checking the time, but it cannot solve
00:23:12for all of them. And all of the data is so intricately intertwined in databases and Excel
00:23:17spreadsheets and all of these things from a client that you're going to have to tie that
00:23:22all together, put it in a manageable, uh, uh, state. You're going to have to understand the business
00:23:27use case for that data to make sure you store it properly, to make sure that you are, are cleaning
00:23:32it properly for that use case. And so if you just asked ChatGPG to clean the data and you'll give
00:23:37it a ton more information about exactly what they're looking for exactly what they want, it's
00:23:41going to have a really tough time doing it. So a lot of that limitation is actually just human error
00:23:46and humans having to be there to help them understand the business context for the business
00:23:51use case, make sure that the data is there. Now I can see some of those pieces AI really helping
00:23:57with
00:23:57you say, Hey, I really want to extrapolate this information. And you say, can we get that with the
00:24:02data we have? They're going to be like, no, we need this other information. So you have to go to
00:24:05the
00:24:05business. You're going to have to say, okay, we need this type of information. Let's start a data
00:24:08collection process. That's stuff that I worked on a lot with. We need to collect this data. Let's
00:24:13set this up. Let's work with our data engineers. Let's work with our database developers. Let's
00:24:16streamline this process, get the data in, and then you could feed it into something like AI and say,
00:24:20okay, we have this data now. Let's loop back around to this business issue. Let's work through it.
00:24:26But again, it's, um, it just can't do everything. It can't. And even if it does do a lot of
00:24:33things
00:24:34and it does a lot of things well, it doesn't do everything perfect. Um, so that's kind of a
00:24:38positive thing. The next thing is human instinct and intuition. It's actually, in my opinion,
00:24:43quite important. Um, I think AI is going to help with this a lot. It's going to help us kind
00:24:48of
00:24:49extrapolate information. Um, and kind of, I use it a lot for things like content creation or generating
00:24:57ideas. I think this is actually going to be a really helpful tool for that. But I think humans,
00:25:02they just have something that AI doesn't, which is just, um, just knowing things, having that
00:25:08intuition in their back, their head that just AI doesn't, you know, it's this, it's this database.
00:25:14It's this pool of information that's pulling from, um, and it's doing its best, but it doesn't
00:25:19understand things, right? It's not like it's understanding what you're saying. It's not coming
00:25:26up with, um, these new ideas, these new things that nobody's ever thought of. Uh, and in some
00:25:33ways it might. Um, but for things like, you know, analytics, there are going to be things where it's
00:25:38going to tell you, Hey, go do this. Um, because this is what we know to be best, but your
00:25:44intuition
00:25:44is telling you something different. And then you have to prompt it. You have to tell it, you have to
00:25:48work with it to be like, no, I don't think that's right because of these things. Um, and I think
00:25:53a lot
00:25:54of teams, not just analytics are going to figure this out. You know, AI is going to be really
00:25:58confident and say, Hey, here's some options. Um, but I think this one's the best. And you're like,
00:26:02that doesn't sound right. That doesn't, that doesn't make logical in intuitive sense. So just
00:26:08being human, having that human instinct, I think is actually really important. Something I don't think
00:26:12AI is going to have anytime soon. The next thing I'm going to read this word for works. I liked
00:26:17how I
00:26:17wrote it. Every company is extremely different. There's no one solution, which is what people
00:26:23think will happen. Every company, uh, every company's data is different. Their needs are
00:26:29different. There'll be hundreds of AI tools for analytics and companies will need multiple
00:26:34companies will need multiple AI tools for different uses, but no one tool will solve all their problems.
00:26:41And there will always be more problems. Um, this is more just, I wrote this mostly because,
00:26:48and again, I like, I, I, I'm trying, I, I'm really trying to grasp AI as a whole. How is
00:26:57it,
00:26:57how it's going to impact the business? And, and that's all my predictions will come up in just a
00:27:01second. And that's where I think a lot of these things come from. But I think that there's not
00:27:07going to ever going to be one solution. I don't think chat GBT is going to grow, grow, grow, grow,
00:27:11grow to where they're, they're the one solution for everything. I think for small businesses,
00:27:17like one to 10, maybe I'm just like 50 people, there will be one or two solutions that they'll
00:27:23use for a lot of their work. But I don't think there's ever going to be something like chat GPT
00:27:29employee who you can just go to as a black box for everything in analytics and get the answer to
00:27:36what's going to connect all your data sources. It's going to do all the data cleaning, understand
00:27:39the business use case. It's going to do everything and just get it done. Realistically in the future,
00:27:44and I'm going to start kind of getting into my predictions. Realistically in the future,
00:27:49every company is going to need something different. There's so many completely unique
00:27:53companies out there. Um, now I run my own company and I know that chat GPT does not solve all
00:27:58my
00:27:58problems. Um, and, and it actually, I use chat GPT quite a bit. I like it. I really do. Um,
00:28:05and,
00:28:05and I see a lot of these other tools that are coming out. I'm like, man,
00:28:08they're really impressive. They're really good. It just doesn't solve all of my problems.
00:28:12I think when you start thinking about it from a business perspective, not just a data analytics
00:28:18perspective, from a business perspective, there's going to be a lot of issues that you run into
00:28:23a lot of issues. Um, and this is kind of the thing I'm positive about. I'm actually, I think this
00:28:29is a
00:28:29good thing for data analysts in general as a whole is that I'm going to take a step back.
00:28:38Businesses don't want managers. They're not going to want managers doing this type of work,
00:28:44using these AI tools. Even if there was a perfect data analyst tool out there,
00:28:49they're not going to want managers doing that. They're going to hire somebody to do that.
00:28:52And most likely if they're a larger team, they're going to want multiple people doing that fact
00:28:57checking, um, fixing prompts. They're going to make sure that the data quality is high,
00:29:03that these data pipelines are wrong or are correct and aren't, aren't wrong. There's going to be a lot
00:29:09of things that you're going to want to double and triple check for in these things. Um, so I mean,
00:29:17I think, I think just as a whole, as a whole, I think, um, every company is very, very, very
00:29:22different,
00:29:23very different. I was going down somewhere with that. Um, but I, I started looking at my predictions
00:29:28and I wanted to get to these now. Okay. Prediction time. Um, this part, I, I, I really, I've been
00:29:36adding to this over the past several months. Every, every so often I'll run into something new. I find
00:29:41and I keep running into these things and I keep adding to it. So here are my predictions. I'm going
00:29:44to start, I don't know how I'm going to talk about it, but I'll start. My first prediction,
00:29:49I think every company is wanting to utilize artificial intelligence and even will, and even
00:29:55more so in the future for analytics. I see that as a huge plus and an upside for data analysts
00:30:00like
00:30:01you and I, I see that as a huge plus because these companies don't have data teams and they now
00:30:06feel
00:30:07they need to create one to stay ahead. So one of my predictions is that a lot of these one
00:30:14man shows,
00:30:16small companies, small businesses are going to want to start using AI. They are not going to want
00:30:21to do it themselves. This technology is useful. If you know what you're doing, it is very difficult
00:30:27to use. In my opinion, still, it's very difficult to use if you don't know what you're doing. Now,
00:30:32when I come into this, I'm coming in from a data analyst perspective. I'm like, wow, this is really
00:30:36intuitive. This is pretty easy to use. That's because I'm very biased towards analytics. I understand
00:30:43it. I've been doing it for several years, right? And so I understand what it's doing. I understand
00:30:48what it's telling me. I understand all these things. So I've been doing it. With that being
00:30:52said, a lot of these business owners, they don't, they don't use data. They don't use these AI tools
00:30:57and they're going to want somebody to come in and do that for them. So I think that overall,
00:31:02that's a really good thing. Um, I actually, I think there's, um,
00:31:09I think there's another point that I was talking about here. Yeah. I'm going to, I'm going to skip.
00:31:12This is a totally different point that I have, but I'm going to kind of merge these two.
00:31:16My number five on here, and I'll go back to two, three, four. Number five is every department in
00:31:20a company is going to start using AI in some way. And we want to optimize their data in that
00:31:24department. Um, so I think we're going to see, and this is, this one I thought is really interesting.
00:31:29And I absolutely am like, I'm like really convinced about this one. We're going to start seeing a lot
00:31:34more specified analytics roles. And what I mean by that is, is right now we have things like
00:31:40healthcare analyst, marketing analyst, financial analyst, right? Those are specific. I think we're
00:31:46going to get even more specific. Um, I think we're going to start going down to like the department
00:31:50level. I I'm talking mostly at like these really large companies. So at a healthcare company,
00:31:55you're not just going to be a healthcare analyst or a healthcare AI analyst is what they might be
00:31:59called in the future. Um, like five, 10 years, who knows? I think you'll get even more specific,
00:32:05right? Within healthcare, you have claims data. You have certain other, uh, breakoffs from that
00:32:11marketing within healthcare. Um, and typically, you know, you have these different healthcare
00:32:16analysts, financial analysts, like I said, I think we're going to have more specific AI focused
00:32:21analytics in all these departments that didn't have analytics before. Um, and people who weren't
00:32:28analysts before are going to become, whether it's called an analyst or it's called an AI prompt,
00:32:35uh, analyst or something like that. Somebody who's working with the data that AI is working with,
00:32:43helping prompt it to get the right answers, troubleshooting its errors. Um, you know,
00:32:49doing these data quality checks, whatever you want to call it. I, I can absolutely see that
00:32:55happening in the future, whether that's 10 or 15 years down the road, just every department's
00:33:00going to want to use AI in some way. AI needs data and every big company has a lot of
00:33:06it.
00:33:07So I think we're going to get, we're going to niche down a lot more. Um, that's kind of like
00:33:11it. I wrote
00:33:12down HR AI analyst. I don't know if that one's going to happen, but that's kind of one of my
00:33:16predictions is I think we're going to get a lot more specific with what we're doing. I don't think
00:33:20the term data analyst, um, in the future will be as broad as it is now. I think you'll see
00:33:27mostly
00:33:28like, um, you know, specific to industries. Um, my next prediction prediction is that I think
00:33:36data freelancing is going to become more popular. Uh, more companies wanting AI, not knowing where
00:33:41to look and the look to consultants and freelancers to come in and help them build their systems and
00:33:46maintain them. The maintenance piece is, I think going to be very important in the future.
00:33:52I think, you know, as I talked about companies are going to want to use AI, a lot of companies
00:33:57are going to want to use AI. Um, I think that freelancing and consulting in AI is going to be
00:34:04huge. I think it's going to be really big. Um, and as these tools, as more and more tools come
00:34:10out,
00:34:10because there's so few tools right now, um, that are in, let me rephrase that. There's so few
00:34:16tools that people have ever used or known about in the AI world, but there's hundreds being built.
00:34:21So I've seen a lot of them, but I've used a handful of all the ones that are out there.
00:34:26So every company is going to be very different. And something I talked about earlier is that
00:34:31we're going to have AI tools for different industries, very specific to their industries
00:34:35to where it has more domain knowledge, more understanding of what's going on. They won't be
00:34:39perfect, but there'll be much more specific to that area. And so you're going to need somebody to
00:34:46maintain those systems, troubleshoot all of those things that we've been talking about.
00:34:50You're going to need somebody. So I think freelancing, I think consulting actually is
00:34:53going to become more popular in the data and AI world. Um, and just a note on that. And I'm,
00:35:00I'm looking at my notes a lot right now because I don't want to like say something that I didn't
00:35:04write down. I want to make sure I'm sticking with what, cause these are like things I've like
00:35:07kind of curated, um, really, and like genuinely thought about the next thing I noted is just that
00:35:12regular people are going to try to do this and fail. Um, as it gets more complex, they're going
00:35:17to hire people with AI and analyst skillset to fill in those gaps. I think the perception that I
00:35:23get is that this is going to make data analysts redundant, or they're going to make data scientists
00:35:29redundant. I've even seen data engineers being redundant, database developers, like anything
00:35:34data related. They're like, it's going to be made redundant. I don't, I don't think that's
00:35:38accurate. What I think is going to be happen is data systems are going to get really good.
00:35:42They're already pretty good. Um, excuse me. Give me one second. Need some water.
00:35:50They're already really good. What I think is going to happen. And I can genuinely, like,
00:35:55I just see it. I can absolutely see this happening, which is a lot of regular people like your neighbor
00:36:02or your HR guy, they're going to try to start using these technologies and they're not going
00:36:08to be as intuitive as they think it is, especially when you're working with complex data. Um, it's
00:36:13going to be, it's going to be more complex. It's going to be less understandable, um, to a lot of
00:36:19people. Like, again, I'm trying to come at it from somebody who isn't in analytics, who really
00:36:24understands creating dashboards and reports and the data and, and the data collection process and all
00:36:30these things. Um, these things are tough. These things are really tough and AI is going to help
00:36:35simplify those things, but these things are going to be tough. They're going to, people are going to
00:36:40have to become data literate, understand the data a lot more. And so people who already have those
00:36:46fundamental skills that have been doing it, that have been studying and creating projects, all these
00:36:50things. I think that, um, is really going to lend itself to transitioning towards these types of jobs
00:36:56that are going to be created. Um, just because people want them, people need them. They see the need for
00:37:03AI
00:37:03in their company. And then they're like, Whoa, I don't know how to do this. Uh, I need somebody to
00:37:08come in who understands this data, who understands the business use cases to help me prompt these
00:37:15things and, and get them all situated, right. And maintain them. That's what I see. My third
00:37:21prediction is I think it's going to lower the barrier of entry for analytics. Um, whether it's good
00:37:26or bad, you can, you know, go both ways, but I think this is going to increase expectations of
00:37:34what the customer or the stakeholder is going to get. So what do I mean by this? I think that
00:37:40I think
00:37:40the first thing is self-explanatory, but I think that the barrier of entry to getting into analytics
00:37:45is going to be somebody who's re just understands the data. Well, I think domain knowledge is going to
00:37:51become way more important than it did in previous years. I've always told you that technical skills
00:37:57are really important and they are. I'm going to talk about that in a little bit. Technical skills
00:38:00are still going to be important, but I think if you really know the domain knowledge of let's say
00:38:07healthcare, I come from healthcare. If you really know the healthcare side of things and you have some
00:38:12data background, you'll be able to get a job faster. So somebody who was a doctor before,
00:38:20somebody who was a nurse before somebody who has, um, you know, worked with on the data side of
00:38:25these things or has firsthand experience with these things are going to be able to transition
00:38:28into analytics easier because they understand the business practicalities that you need to know.
00:38:34They understand, I don't know if practicalities was the right word, uh, the business practices.
00:38:38If they understand the business practices, they understand how the data is collected, how the data
00:38:41flows. And then they start using these tools. They start learning the data side of things. I think the,
00:38:48I think the barrier of entry for those people transitioning in is going to be a little bit
00:38:52lower because of these AI tools. Um, the expectations of the customer are going to increase.
00:38:59This one is, I think is, has always been true. So I don't think there's anything new. I'm not saying
00:39:04anything revolutionary here, but here's what I will say. When you start using these AI tools,
00:39:10the expectation is that the quality is going to go up quite a bit. They're paying for this technology
00:39:15that is supposed to be state of the art best that you can buy. And their, their expectations are
00:39:20going to go up. So you as a data analyst out there, you know, we're coming into a company,
00:39:24they're going to expect that, you know, how to use these things really well to get exactly what
00:39:28they need out of them. I think that is my part of my prediction in this. And this is a
00:39:33sub note
00:39:34is I think this is actually going to cause a lot of frustration in the data world,
00:39:40because I think so people are going to have such high expectations for AI in general,
00:39:45and it's not going to get them a hundred percent there. Um, it just, as of right now and where
00:39:52I
00:39:52see in the next five years, 10 years, it's just not going to be perfect. And there's a long way
00:39:56to
00:39:56go, a long way to go for it to be really, really good where it doesn't need as much oversight
00:40:03and
00:40:03hands-on experience or hands-on, um, oversight is what I was saying that you wouldn't need something
00:40:09like a data analyst. It needs a lot of work. Um, and even then our company is going to trust
00:40:15that
00:40:15it's really tough to say. Um, and so the expectations are going to get a lot higher.
00:40:21The next one, I think that my number four is knowing analytics and the tools are become
00:40:28very important. Um, as AI starts utilizing tools, if something goes wrong, you're going to need somebody
00:40:34who knows the systems and software well to troubleshoot and get it right. Now, this is
00:40:38happening already with things like auto GPT, baby GPT, using these tools, um, things like plugins
00:40:45within chat GPT, which I think are going to be even more so and things like GitHub, uh, not GitHub
00:40:53copilot, Microsoft copilot 365. That's going to be integrating chat GPT into a lot of these things.
00:40:58Uh, and a lot of their Microsoft applications, what happens when something doesn't go right?
00:41:04You need somebody who knows these tools well. So what I foresee is people are going to hire people
00:41:09who already know these tools to come in and use them, monitor them, use them with AI. So if they're
00:41:16wanting a PI, a power BI visualization specialist, um, I, or let's say they're, you know, uh, very common
00:41:24job right now as a data visualization specialist, data visual, data visualization specialists, in my
00:41:30opinion, are some of the easier jobs to automate in the data analyst sphere. Cause data visualization
00:41:39from everything that I've seen is something that, um, AI gets pretty right or gets it better than others.
00:41:47Things like data cleaning, it gets wrong, um, quite a bit. It doesn't understand all the business,
00:41:52uh, side of things. So having that domain knowledge, but data visualization, I feel like it gets,
00:41:57it gets right more often than not. Now you still need to prompt it for other things.
00:42:01You still need to, um, create different types of visualizations for different things. So I don't
00:42:06think that job entirely is going to go away. I just will, I do think that that job specifically
00:42:12is more susceptible to being automated in my opinion. Um, so I think knowing these tools though
00:42:20is still going to be extremely important because as you get into it more on a base level, you could
00:42:28come in who somebody who knows none of these technologies and they could kind of get by, I
00:42:32think. Right. Which is impressive with AI. They could kind of get by doing a little bit of data
00:42:37analysis, but as you get into the more complex things and you're starting to integrate different
00:42:42tools, like let's say it's, you know, writing all your Python for you. It's trying to create all your
00:42:45visualizations for you in a different tool. It's creating data pipelines. It's, it's, you know,
00:42:50cleaning data. You don't want just some random guy doing that. That is still very specialized.
00:42:56So you knowing these tools, knowing how to use them, how all these pieces fit together and helping
00:43:02these AI systems do that properly, I think is actually going to be really important. I don't
00:43:07actually see that happening to that level. Like just somebody kind of knowing the tools
00:43:11and kind of piecing everything together for AI to do it. I don't see that exactly happening
00:43:16for several years. Even then I, you know, it's, it's tough for me to imagine AI being that
00:43:23good where you just need somebody to supervise. I think it's still going to require human intervention
00:43:31going in, helping with this work, helping prompt them, help them understand things. Um, again,
00:43:36I just don't see that happening and I'm going to skip to number six. Um, and I kind of,
00:43:42I kind of mentioned this already. Um, but again, this is like my thoughts over the past
00:43:46several like months. Number six is I think more and more companies are going to try to
00:43:51use AI on their own and realize it's not that easy. It's not just a plug and play and
00:43:56you get everything you need. You need good data. You need good pipelines in place. You need
00:44:00someone who knows what they're doing. As this continues to happen, more and more companies
00:44:03will turn to data professionals to come in and help maintain these solutions that they
00:44:07thought they could do themselves. Um, it's just a general sentiment that I've seen online
00:44:12is that people are going to do everything themselves or they're going to have AI come
00:44:17in and do everything for them. I just, um, I personally don't see that happening. I think
00:44:22that a lot of the, and again, a lot of it comes back to like time and accuracy managers, small
00:44:32business owners, um, people like me, like I can't do everything myself. I can't like I'm
00:44:39growing my business and there are things that even if I wanted to, I couldn't automate it to
00:44:44be done. I need to have someone, even if I have them use an AI tool, I need them to
00:44:48come
00:44:48in and use it for me. Um, at least right now, right? Let's imagine. And, you know, and next
00:44:53thing I'm going to talk about is five years, 10 years, 15 years, one of my predictions,
00:44:58but you know, as of right now, I would need to hire them to come in and kind of manage
00:45:03that
00:45:03system. I could, I just, I'm not going to do it myself. That's ridiculous. It does not make
00:45:07sense for me to do that for my company and does not make sense for other people to do that
00:45:10in their
00:45:10company. Um, and so as long as AI needs some oversight and it will need it for a long time,
00:45:17that is, I just, I'm sure of that. As long as it needs oversight, there's a lot of work to
00:45:22be done
00:45:22and that's not going to be something that people just like regular people who have no idea what
00:45:26they're doing are just going to figure out. Um, unless they study it a lot. It's just, I don't,
00:45:32I don't see that happening. Um, so this is where I'm going to get, uh, I'm going to talk about
00:45:42predictions five years, 10 years in a little bit, but I want to kind of bring us all back to
00:45:47about
00:45:47like five or six months ago. Um, many of you guys may not know this. How much time do I
00:45:52have?
00:45:54Okay. Many of you guys may not know this. I'm, I'm building a company right now.
00:45:58My, my consulting company, right? In analytics. I'm also building a website called analystbuilder.com.
00:46:06That's going to have, uh, you know, courses and a lot of analyst focused things on there.
00:46:10And about five months ago, I literally like stopped and, and as all these things were developing and I was
00:46:16like, Oh my gosh, I don't know if I'm going to have a job in two years. Very, very genuinely.
00:46:25Like I,
00:46:27I, for about almost a full week, I was just like, I don't want to say depressed. I don't want
00:46:34to say
00:46:34that's the word, but I very much was like, Oh my gosh, I was having like an existential crisis. Like,
00:46:39what do I do? And I really more or less started worrying about my audience. I was like,
00:46:49I was like, does this mean everyone who's like following my channel? Does this mean,
00:46:54you know, there's no future for them? I mean, I really, I had this like really like serious moment
00:47:00where I was just like, I don't know if this is, I don't know what's going to happen. I really
00:47:04don't.
00:47:04Cause things were advancing so quickly. I couldn't keep up. And it felt like at every turn,
00:47:10things were getting better and things were getting better.
00:47:14After that week I started, I was like, okay, I got to stop sitting on my butt and feeling sorry
00:47:20for
00:47:20myself. I need to like really dive into this and seeing how viable this is. Cause I, I, I believe
00:47:26I have a duty to you guys to inform you the best I possibly can. Like if I thought analytics
00:47:32was going
00:47:32away forever and they were all going to be automated, I would bring you the honest truth
00:47:37as I dug into it more. And I dug into it a lot as somebody who works, who's worked on
00:47:43kind of every
00:47:44process and analytics from data collection to data cleaning, to business side, developer side,
00:47:51visualization. I feel like I've worked on a lot of different areas. You know, the more I started
00:47:57using it and trying it out in these different areas, I'm like, man, this is extremely impressive.
00:48:02But I can already see a lot of cracks, a lot of cracks and something that I kept finding over
00:48:11and
00:48:12over and over again over the past four or five months. I really been diving deep into it is that
00:48:19AI is just not, I don't see it taking over. I don't see it. Let me rephrase that. I not
00:48:25taking
00:48:26over. I don't think that's the right word. I don't see it replacing data analysts fully.
00:48:32I see it being a huge tool and asset, but I also see a lot, you know, I see those
00:48:39fears and concerns
00:48:40because I was in that exact place like four months ago, five months ago. I was in the exact same
00:48:44place.
00:48:45Very much worried. I was like, not just for my, not just for me. It was more like,
00:48:50I mean, I feel like I'm a teacher on YouTube. That's how I feel. I feel like I teach people
00:48:57data analytics to help go out there and get a job that they want and that they love and that
00:49:01they'll enjoy doing. I'm like, is everything I've built as all the people that I've been teaching,
00:49:07are they just, is there no future for them? And I, from my personal opinion, as of right now,
00:49:14on March 8th of 2023, I really don't see that happening. I don't think, I don't think we're
00:49:22looking at mass layoffs soon. What I genuinely think is going to happen is it's going to be a
00:49:29somewhat slow rollout, even though the technology is advancing very quickly. I think the rollout
00:49:34of it being integrated into all these companies is actually going to be a lot slower than you'd
00:49:38imagine. And I think as, um, they roll out a lot of these artificial intelligence, chat, GPT,
00:49:46LLMs, whatever you want to say, as they roll these out into companies, we're going to find a lot more
00:49:50issues. We've been using this personally on a personal scale. Once you start integrating into
00:49:57companies, you're going to find a lot more issues, um, with privacy, ethical concerns, errors,
00:50:05hallucinations, which have not gone away. And I haven't even touched on, but that's part of the
00:50:09mistakes. It still makes hallucinations. Um, I think that we're going to find a ton of issues,
00:50:16a ton. Um, and over time, those will get better, but my prediction in the next five years from now
00:50:26till five years. So we're looking at 2028. I see AI systems starting to be integrated at most
00:50:32companies within the next five years, uh, or integrated in companies in the next five years.
00:50:37I don't see, um, I don't see many data analyst jobs being replaced fully. I think that there may
00:50:48be a small slight slowdown in hiring an AI, um, in data analysts in the future. Now there are people
00:50:54that retire. There are people that come in. There's already, there has been for the past six years since
00:50:59I've been in data analytics. There's already been a tough barrier of entry for entry level who have
00:51:04no experience. I think that'll always be there. Um, so I don't see any different in the next five
00:51:09years. What I will say is data analysts who have experience entry level, mid-level, senior level,
00:51:15you're going to find more opportunities. And I think, um, more companies hiring for data
00:51:22professionals with these AI systems in the next five years. I think we will see an, uh, an increase in
00:51:28demand, uh, for that with that as a lot of opportunity from people who aren't data focused,
00:51:33people who were, like I said, a nurse who wants to break, break into here. I think it's going to
00:51:39make
00:51:39it easier. So if that's you, that's fantastic. But if you're coming out of something like high school
00:51:47and you don't have no experience, you have no experience in any job whatsoever, I think it's
00:51:52going to be tougher for you. Um, so people with domain experience, people who have already worked
00:51:58jobs that wasn't data focused, but understand the business side of things, I think it's going to be
00:52:02even better. So that's five years. Now let's talk about 10 years, 15 years. I'm going to kind of lump
00:52:07that just into the future. Like what does the future of analytics look like? I see AI being integrated
00:52:12into a lot of processes and analytics, just like I think most of them. I think there will always be
00:52:19need for humans to, to have their hands in the analytics side of things. I think analytics is here
00:52:26to stay. I don't see analytics going anywhere. I really don't see analytics being automated fully
00:52:31ever. Um, at least not in the foreseeable future. I couldn't imagine in the next like 10 or 15 years.
00:52:38I
00:52:38just can't, I can't see it, especially with all the issues that I know are going to arise in the
00:52:45next
00:52:45five years with AI. There's going to be a lot of companies that just straight up ban it.
00:52:49Another prediction. Um, I think there's gonna be a lot of companies that, that ban a lot of AI tools
00:52:54for a lot of different reasons. Um, and I don't think, um, I don't think, you know, just to preface
00:53:05this, that in the next 10 or 15 years that we'll see, uh, an extremely dramatic layoff. Like we think
00:53:12about in the news where you're thinking Google is just going to be automated. Um, entire departments
00:53:17are going to go away. I, it just, it seems in my opinion right now, somewhat unfathomable,
00:53:25um, because I just know how I've been doing so much research on these systems. I just feel like
00:53:30I know they have so many downsides. They have a lot of upsides. There's a lot, a lot, a lot
00:53:35of
00:53:35opportunity out there, but there's a lot of downsides as well. I think a lot of companies
00:53:39will see that once they start using them, they're gonna be like, wow, this is amazing. But
00:53:43I'm already seeing some of the cracks and the downsides of using AI. That's, that's kind
00:53:48of what I think. Um, so I'm going to, I'm going to just a second go into what I think
00:53:57you should be doing right now as somebody who's breaking into analytics, as somebody who is
00:54:03trying to get an entry-level analyst, analyst job. Before I do that, I just want to say, um,
00:54:10I've been in the headspace in the past of like doomsday prepper. This is like the end
00:54:16of analytics, robotsy, um, you know, mass layoffs and mass. I've kind of went down that rabbit
00:54:22hole for a little bit all the way to, um, you know, AI can't do what humans do. It's
00:54:29just, they can't do it. They're just robots, uh, you know, train or whatever. They can't
00:54:33do what we do. Bada, bada, bada. I've kind of been on both sides at different times. Now
00:54:37I'm kind of in the middle where I think that there are some very real concerns for people
00:54:43who are trying to break into data analytics. And I think those concerns are extremely valid.
00:54:47And I hope I've kind of, you know, validated those fears because I don't want you to feel
00:54:51like you're the only one thinking about these things. I also see a lot of positive things
00:54:56about AI. It's going to speed our work up a lot. It's going to allow us to do better
00:54:59work, faster work, um, and get more done. Um, and I think that's a fantastic thing. I, in,
00:55:06in general, I think pro I am AI, I'm pro AI. There are areas where I'm like, if this were
00:55:13to go South, if AI started doing everything, um, somehow, or it started, um, you know, like
00:55:22in the worst case scenario, it starts like, you know, taking over your systems or, you
00:55:28know, whatever, like Skynet or something, right? Or Terminator. I, you know, however
00:55:34this goes, I could also see that being very bad. But again, I, I've been talking about,
00:55:41I've been talking for like, I feel like an hour. I feel overall positive about it. I am
00:55:48skeptical. I, I'm hesitant as well as others, right? I'm just trying to, you know, I'm trying
00:55:56to be as honest as I can. That's my genuine take on it. So what would I do if I
00:56:02was just
00:56:02starting out today? Now I, I have what, six, almost seven years of experience now working
00:56:10in healthcare, IT, consulting. I feel like I have a lot of experience. So I feel like for
00:56:16me, I have a lot of these tools. I have a lot of these skills. I have a good, um,
00:56:21a
00:56:21good reputation. I think for someone like me, it's not going to be crazy hard. A lot
00:56:27of other people who have experience like I do, who have experience in analytics and have
00:56:33for several years, I think most of you guys are going to be okay. I genuinely do. I don't
00:56:37see a huge, um, mass layoffs for senior level, um, even mid-level analysts. I think the entry
00:56:45level analyst is going to be the new mid-level. So here's what I would say is that I would
00:56:50do. And this is my personal belief. You don't have to do this is what I would be doing. If
00:56:55I was starting from scratch, I was like, I just want to break into analytics at some
00:56:59point in the future. I would learn the basics first, because even with all of these, uh,
00:57:08specifically, let me go back. What's it called? Microsoft's Copilot 365. ChatGPT is going
00:57:12to be integrated into things like Excel very soon. Something that data analysts work a lot
00:57:17in. I would still learn the concepts for using Excel. And the reason for that is because when
00:57:27you start using ChatGPT for Excel, it's going to start popping a lot of these functions, a
00:57:32lot of these pivot tables and visualizations, and you should know what it's actually doing.
00:57:38I am not a huge believer that you should just let it do what it does and blindly believe
00:57:43it. Um, so I would still highly, highly, highly recommend learning the basics of Excel, SQL,
00:57:50Tableau, Power BI, Python. And, um, I think that's really, really important. I think that
00:57:56Python, and that's a lot of what ChatGPT, uh, is built on and used for and in data analytics
00:58:03and stuff like that. Python is super popular. Um, I would learn Python probably, um, you
00:58:10know, I would learn that one maybe sooner than I have said in the past where I'm like,
00:58:15you don't really need to know Python. I personally think Python is going to be, or stick around
00:58:19and they're going to be different programming languages, um, that are built off of that
00:58:22specifically for AI, like an AI language that, um, one of my, uh, buddies just sent me called
00:58:28Mojo, which is literally built to be read like Python, but it's for AI languages. So something
00:58:35like Mojo, those are things that you, if you know the basics of Python, you're going to
00:58:38be able to pick that up quick and learn that. So I still highly, highly, highly recommend
00:58:42learning the basics, even up to intermediate and advanced levels. If you can, the reason
00:58:48for that is after you start, after you know those things, I do believe that you need to
00:58:53start learning AI. You don't need to be an expert in it, but you need to start dabbling
00:59:00in it at a minimum. I think that is going to be extremely important in the future. So
00:59:06I would start dabbling with chat GPT, start trying to do some data analysis with it, start
00:59:11testing it. Um, you know, that will only help you learn, right? You're going to learn, but
00:59:17you're also going to make yourself more job ready for the future. Five years down the road,
00:59:21I think a lot of data analysts are going to be using AI on the daily. So I'm kind of
00:59:26adding that to my like a bucket of things you need to learn, learn AI, start with chat GPT.
00:59:32I think it's like one of the simpler ones. Then, you know, this, um, co-pilot is going
00:59:36to come out not too long. So you can start messing around with it in like PowerPoint and Excel
00:59:41and, and you know, different tools and then plugins with chat GPT into your, into these tools
00:59:46and start messing around with those things. Again, this is May 8th of 2023. In six months,
00:59:54chat GPT could have plugins into Power BI, which I'm sure, uh, they're already going to do that
00:59:59through Microsoft, but it could have plugins into Tableau, into Looker, into, um, any tool that
01:00:06you're using. Start learning the basics now, start learning chat GPT right after or AI. Um,
01:00:16and then, oh yeah. So I basically said, learn those skills, learn how to integrate AI into them
01:00:20and then learn how to solve business problems. Um, I think understanding the business side of
01:00:27things, understanding your client's needs, your client's wants, your stakeholder needs and wants,
01:00:31and translating that into, um, actionable insights. You cannot go wrong. And that's tough to do when
01:00:39you're first starting out. Um, find business use cases, you know, find projects online. Um, I have
01:00:47tons of, so something that I would do, I have tons of projects online that I've done for free on
01:00:50my
01:00:51YouTube channel, right? You can take those and try to replicate that using AI. Try to do what I did
01:00:58in
01:00:58those projects. And you can use AI to help you build that, do it faster than I did it, right?
01:01:05That is something that I think is a very real use case. And I think in the next two to
01:01:10three years,
01:01:10if you're doing those things, you are going to really set yourself up well. Again, I don't, I, I,
01:01:18I, maybe I'm just a sucker. I'm just really a positive person. I just don't see analytics
01:01:25going anywhere. I think that we're going to have more data and more opportunities with the data
01:01:33that's being generated in the future. AI is going to help with that. I think it's going to be a
01:01:38huge
01:01:38tool, a transformative tool, something that, you know, I couldn't have comprehended even 10 years
01:01:43ago thinking about like, like college. Like it's, it's a huge transformative tool. It's going to
01:01:49transform a lot of things, but it's still, in my opinion, a tool. It's not somebody who's going to come
01:01:53in and do everything for you. So that's, that's what I'm thinking. Now, normally during these,
01:02:02um, these episodes, these long, long episodes, I do two things at the very end. I do a question of
01:02:09the week, which, um, I'm not going to do. I'm not going to do because I think this is a
01:02:15really serious
01:02:15topic. I'm not trying to, um, I've run, I've ran a long time, so I'm not going to do a
01:02:21question of the
01:02:21week this week. What I will do is a vegetable of the week. Um, and the vegetable of the week
01:02:26is just
01:02:27saying, Hey, I washed all the way through and it doesn't mean you necessarily agreed with me.
01:02:34It just means you're taking this seriously. Uh, so I'm going to reuse a vegetable. And because this
01:02:45is so serious, in my opinion, it's such a serious topic, I'm going to reuse one that, um, has a
01:02:50real
01:02:53close part, uh, uh, holds a soft spot in my heart, which is from one of the very first episodes
01:02:59I ever
01:02:59did, um, for the Alex, the analyst show. And then the vegetable of the week is going to be jalapeno.
01:03:05So if you watched all the way through and you were, you could be concerned about AI, you could be
01:03:12hopeful
01:03:12about AI, you could just be getting into analytics. You could just be, um, you know, looking for a senior
01:03:18position in analytics. You're watching this because, you know, either you're, you're nervous, you're
01:03:23excited, or you're not sure what to think. And that's what this is all about. And so if you stayed
01:03:29all the way to the end, put jalapenos in the comment section below, I will know that you were taking
01:03:34this
01:03:34very seriously. You care about your future. If you don't put in a jalapeno in there, I'm going to assume
01:03:40you
01:03:40don't care. But seriously, um, this is a heavy topic. I, I, I genuinely, I put weight, not way too
01:03:50much thought. I put a lot of thought, um, just because I care so much about it. This topic means
01:03:57so much to me. The, you guys mean so much to me. I wanted to make sure I was bringing
01:04:01you good
01:04:01information. I hope I brought you good thoughts that maybe you haven't thought about before. Um,
01:04:06good ideas that you haven't thought about before. And I hope, um, cause I know there's a lot of
01:04:11people out there that are really worried. I hope that I put some people at ease in my heart. I
01:04:16am
01:04:16at ease. I am not crazy worried. Um, just because I feel like I have a pretty good understanding for
01:04:24how companies think and work and operate. Uh, you know, I, I feel like I have a pretty good instinct
01:04:31on this, um, just on, on how they're going to respond to AI, how it's actually going to go with
01:04:35implementing these AI systems. So I feel like, uh, I hope for the most part, um, AI I see is
01:04:44a very big
01:04:45tool and I'm hoping I put some people at ease. It's a very serious topic. Um, it's people's
01:04:50livelihoods. It's people's futures. It's crazy important. I can't under, I can't, I can't say this
01:04:57enough. It, you know, it just, this technology has the ability to transform lives for the better,
01:05:01for the worse. Um, and for the majority, I think it's going to transform it for the better.
01:05:07I'm not, I'm not going to, I don't know if I refuse to believe it or refuse to think it,
01:05:14but I'm not going to succumb to, um, kind of the doomsday prepper, um, end of the world end of,
01:05:21uh, data professionals in general. And there are people out there like that. I just think they
01:05:27maybe haven't been in the industry long enough or understand it enough, or they're just crazy
01:05:32negative people. I don't know. It's hard to say. Um, you know, I kept saying the date on here
01:05:39and I kept saying that because I want to come back to this in one year. Like I'm making a,
01:05:44a note. I
01:05:44already have the note on my calendar for when this launches, which should be next Tuesday. Let me see.
01:05:52Why is my thing not here? What the heck? Here we go. Uh, next Tuesday is the 16th. So on,
01:06:00um,
01:06:00on May 16th of 2024, I'm going to come back and see how accurate it was. Um,
01:06:08a lot of things are going to change. A lot of things are going to change. I just, I have
01:06:12a strong
01:06:13hunch and all of my research has pointed to that it's going to mostly be for the better.
01:06:19That's what everything's telling me. I hope I'm right. I really hope I'm right. If I'm wrong and
01:06:24in five years, you know, analytics is completely gone. You guys can be mad at me and I want you
01:06:29to
01:06:29be mad at me. But with all the information I have now, I'm giving you the best. I'm giving you
01:06:33the
01:06:33best I got. I really hope, I really hope if I'm wrong, you guys can forgive me. I hope if
01:06:40I'm right,
01:06:41you guys can, um, celebrate with me. But this is a, again, now I'm just rambling. You guys can go.
01:06:49You already typed jalapeno down below. I'm just going to hang out for a little bit.
01:06:53This is a long episode. Anyways, nobody cares. Nobody cares if it goes an extra five minutes.
01:06:59Just a really, I don't know. It's really been weighing on me too. It's really been weighing on
01:07:05me. It's really been weighing on me. Now, if you're staying, if you're sticking around to the
01:07:10end, I'm going to cut you in on some secrets. Um, and some other things that I've been thinking
01:07:14about, you know, I'm starting this company or starting this, this platform and website called
01:07:19analyst builder. It's not released yet. Um, it'll be ready in like a month or two. I've been working
01:07:23on it for over a year and it's for data analysts to learn data analysis, to practice data analysis.
01:07:32And I'm going to grow the website to be something incredible. I think it's going to be one of the
01:07:36best data analyst websites that is out there. What like legitimately, it's already incredible.
01:07:42Um, as is, and it's simple. It hasn't even gotten like to the stuff that I want to get on
01:07:46there.
01:07:48And four or five months ago, when I had that like moment of like, Oh my gosh, it's over.
01:07:53I was like, AI one, there's no use in me building this anymore. I should stop,
01:07:59tell the developers to stop building. I legitimately had that thought. And
01:08:04the more I looked into the more I'm like, the more I'm like, this needs to happen.
01:08:09This is going to be more important. There's going to be, this will be my platform
01:08:12to help ushering kind of the new data analysts of the future where I'm going to start teaching.
01:08:17I mean, I'm going to, I'm diving deep right now into AI. And so, you know, one of my goals
01:08:23is to integrate AI into the platform, do courses on AI, teach people how to use AI as well as
01:08:29on
01:08:29YouTube for free, of course. Cause I, I, you know, this website, of course, will be some things
01:08:34will be paid for, but you know, these are the things that people are going to need to learn
01:08:40and someone's got to teach them and universities are not going to do it anymore. I random rant,
01:08:47but with chat GPT, AI, LLMs, all these things, the need for higher education is going to go down
01:08:58substantially in certain areas. I just, I know it in 15 years. I think it'll be like you went to
01:09:04college. That's pretty cool. Nobody does call it. None of my friends went to college. It's going to
01:09:08be like that. It's that, it's that transformative. And so, you know, I don't think waiting for
01:09:16education to catch up is the way to go. I think platforms like mine are going to be,
01:09:20I don't want to say the future, but it's, you know, for people who want to break into analytics,
01:09:24my goal is to be on the front lines and creating content on YouTube and an analyst builder to
01:09:31kind of keep people up with technology and what's out there and what's being done.
01:09:37And that's, that's, you know, that's where my heart is. My heart is still in YouTube. It is still
01:09:43very much in my audience and my community. So I feel like I owe it to them. You guys have
01:09:48given me so much. I feel like I legitimately, like, I feel like I owe it to you guys to
01:09:52keep
01:09:52going. Um, and that really pushed me through to be like, okay, if that's what I'm thinking,
01:09:59I need to get out of this funk. I need to really dive into this and see like what, if
01:10:04this is all
01:10:04true, like you'll see claims on like Twitter and LinkedIn, like AI is ruining data science is
01:10:09ruining analytics. Like there's no jobs. Don't do it. The more I looked in, I was like, that's
01:10:14just not what I found. And so I'm really happy that I kind of got out of that funk. I'm
01:10:19still
01:10:19here for you guys. I am still going to be creating content. Now I'm just looking at myself. I'm still
01:10:26going to be creating content. I'm still going to be creating courses. I plan on staying as up to date
01:10:31on this technology as I can so that I can then teach you guys. Um, and I want you guys,
01:10:37my audience, my community, I want you guys to be the first ones learning these things that you can
01:10:42then go out there and get these jobs because they're going to be jobs. There's an, it's not
01:10:47going away. I've really, I'm just so certain about it. With that being said, I've talked for long
01:10:53enough. I am now getting headache. And I also have blood drawn this morning for life insurance policy.
01:10:58Um, I, I had some blood drawn this morning. So now I'm a little lightheaded. So I've been sipping
01:11:02on water. So with that being said, um, I hope, I hope that I brought some positivity or positive
01:11:13thoughts on it around. Um, it's just a heavy topic. It's a really heavy topic and I can't,
01:11:19I could try to make it fun, but then I just, I'd feel disingenuine because this is a very heavy
01:11:23topic
01:11:24in my opinion. So, um, yeah, that's how I'm going to end it. Just, it's a heavy topic. Um, but
01:11:32I'm,
01:11:33I'm overall positive. I keep saying it, but you know, I just keep thinking about it. I'm overall
01:11:37positive. Thank you guys for joining. I do appreciate it. I really do. If you're out there
01:11:43listening still, I mean, you guys are the real heroes. You guys are the real heroes. Um, so with that
01:11:49being said, have a good week, I'll see you on March, May 16th when this lunches, I'll see you then.
01:11:59All right. Well, take it easy. Uh, have a good one. Enjoy your day. Give your significant other,
01:12:09family member, friend, child, a hug. Not from me. That'd be weird. Uh, just from you.
01:12:17If you really want it to come from me, you could, don't do that. Don't do that. All right. No,
01:12:26for real, I'm out of here now. Have a good one. Goodbye. And I will see you in the next
01:12:30video.

Recommended