• 25 minutes ago
Is data analytics disappearing in the age of AI, or changing into something more demanding? Alex The Analyst explores what the next three to five years could mean for data analysts, from fewer pure reporting roles to growing demand for business knowledge, data modeling, semantic layers, and end-to-end ownership of data processes. The discussion also looks at how hiring managers are weighing analytical thinking alongside coding skills, and how AI is reshaping self-service analytics and the modern data stack.

For anyone planning a data analyst career or preparing for analytics jobs, this practical career analysis highlights skills to build: SQL, Python, AI workflows, cloud platforms, and tools such as Databricks, Snowflake, AWS, and Azure. Discover insights on the future of data analytics, AI and data analysis, data analyst skills, business intelligence careers, and modern data tools.
Transcript
00:00Hello, everybody. In this video, we're going to be talking about the future of data analytics.
00:04And I get asked about this topic all the time. In the world of AI that we now live in,
00:09where does data analytics play a part? How is it changing? What are things that I should be doing
00:14to stay relevant and competitive in this job market? I actually wrote and released last week
00:18an entire blog on this topic. I tried to write just enough, not too much, on some of the more
00:26important things that I think we should be discussing, like how data analytics is changing,
00:30the top skills that analysts need over the next three to five years, what hiring managers and
00:35analytics are actually looking for right now, and just how the modern data stack is changing.
00:39These are all things that we are going to discuss and talk about in this video. Now, in my blog
00:44post,
00:44as well as in this video, this is my opinion. And I have talked to a lot of other industry
00:49leaders who
00:49are either CEOs of small startups, or they are, you know, directors of analytics at different tech
00:54companies, and I've gotten their feedback as well. And it's kind of shaped my opinion, you know,
00:59over the last several years. So to answer the elephant in the room, or kind of the big question
01:03that's on everyone's mind, is data analytics actually dying? And in my opinion, from everything
01:08that I've seen, no, it's not dying, but it is changing. And it's been changing for several years.
01:13So here's how data analytics is actually changing. And what I kind of predict will happen even more
01:18within the next three to five years. The first thing is, I think we're going to see less people
01:23doing pure reporting. And I've worked with people who are literally BI dashboard analysts,
01:28they just build dashboards. And although that's really important, because I still see value in
01:34dashboards, I just don't think you're going to see as many dedicated jobs that anymore, because AI has
01:39gotten really good at purely building dashboards, once you've done a lot of the work kind of before
01:46it. And so if you're able to define everything and explain everything, it can actually build a fairly
01:50decent dashboard. It's not going to be perfect. And I've have lots of videos. And I've done a lot
01:55of actual work doing this does a pretty good job, though. And so I don't think we're going to see
01:59as many pure reporting analyst roles. The next one, and I think we're already seeing this is there's a
02:05lot bigger demand for people who really understand the industry and understand the business and the
02:10data model. Now, even five or 10 years ago, you could get a job because you just knew a tool.
02:14Hey, I know how to use Azure. I know how to use Tableau. I know how to use Power BI.
02:19And those
02:19tool specific roles didn't really care, especially if you knew the business super well. It of course
02:25was a big help, but they would hire you regardless. They're just like, hey, we need someone who knows
02:29Power BI. I think that version of roles that we're looking at are kind of going away. The expectation
02:35kind of going forward is that you understand how the business works. You understand the clients,
02:39the customers, the needs, the wants. It's a lot more of that domain knowledge that they're
02:44really looking for. Companies need somebody who can actually build out these semantic layers,
02:49these business models, the data models, so that when they are kind of integrating everything in
02:54whatever system they're using, it all works properly and is all well-defined. And that takes
03:00someone who understands the business well. The next thing that's changing, and we are already seeing
03:04this happen kind of across the board, is companies are trying to streamline things down to smaller,
03:10more senior teams. I would say this is especially prevalent at smaller companies, especially at
03:16startups where they're trying to like really make use of all the money that they have, or as larger
03:21tech companies, they can afford to hire some, you know, junior people. Now, although we are already
03:25seeing this, I think this is going to be an even bigger issue. And I call it an issue because
03:29it really disrupts this process that we've had within, you know, tech and analytics for a long time,
03:37which is you have a senior, you bring on a junior, and that junior trains under this senior analyst
03:42to really learn a lot of things and learn how things actually get done. The issue with this
03:47pipeline, though, is that now we're not going to have as many juniors starting out working with
03:52those seniors. There's going to be this expectation that these people who are just getting into the
03:56industry can do a lot of these more senior level things because of AI. And the issue with this is,
04:02is that if you don't actually have hands-on work, and you're relying a lot on AI for a lot
04:08of these
04:08things, you don't really learn it that well. You're kind of handicapped in the fact that maybe
04:13you aren't going to learn the business as well. Maybe you're not going to understand the domain
04:16as well, because there's just this expectation that you kind of already know it, or you have the
04:21ability to learn it by using some type of AI tool. So although it makes sense for companies to do
04:27this, I don't like that it's happening, but it is happening. My advice for people who are just
04:32getting into analytics is to really find that differentiating thing. I would not focus purely
04:37on tools. I wouldn't be the guy who, hey, I know SQL, because now a lot of people know SQL.
04:43You still have to know SQL. That's just an example. But that cannot be your thing anymore. You have to
04:49be the guy or girl who knows how to actually get things done with the tools. Hey, that's the person
04:55who is able to do X, Y, and Z. That's the person who knows this industry really well. It's more
05:00about
05:00domain knowledge than it is about tool knowledge these days. Again, with that being said, you still
05:06have to know the tools. You can't just get any job and then learn those tools on the job. For
05:12the most
05:13part, you're going to be expected to know those tools. So still learn, you know, SQL and Python
05:18and Databricks and AWS and Azure and all these other things. But don't think that that is going to be
05:23the
05:23thing that gets you the job. Now let's take a look at these skills that I think every single
05:27analyst needs to know over the next three to five years. Ones that I think are going to become
05:31even more important if they aren't already important today. Believe it or not, the first
05:35one is still SQL. You still need to know it. You still have to get in and use data. Now
05:40we're seeing
05:41tools like Databricks. I'm going to get into kind of some of the future tools I think you should be
05:45learning later. But let's just take Databricks for example. Databricks has something called Databricks
05:50one where a business user can go in and they can query the data and they can look at graphs
05:55and
05:55visualizations all just like an AI integration. They don't ever have to write SQL. But that's
06:00just a business user. And they're going to have all sorts of questions that are going to come from
06:05this because they're going to say, hey, these numbers don't match. Or, you know, I looked at
06:09this and that doesn't make sense because of this. Can you dive into this? As the data analyst on the
06:13team, I need you to give some clarity to this information. And so you can use AI within Databricks
06:19and absolutely you should. But I've been using it enough to know that it doesn't always understand
06:24what I want. Because I personally know SQL really well. When I am in Databricks and I'm prompting AI
06:30and it's hitting off all my databases and it's doing joins and window functions and all these things,
06:35as it's writing it out and as I'm looking at it, I know almost immediately if it's right or wrong.
06:40Like almost immediately. I can say, yeah, okay, I see what it did. It didn't do it right. Let me
06:44prompt
06:44it to change it. But most people who don't know SQL that well are just going to accept it. And
06:50they're going to take whatever output is given. And so knowing SQL well and really understanding
06:55how it works is still an incredibly good skill to have. The next skill to know is data modeling
07:00and semantic layers. I'm kind of bunching those together because they do intertwine a lot.
07:05Creating these definitions for a business is really important, especially when you're working
07:09with lots of different departments who all kind of work with somewhat of the same data,
07:14data, but just in different ways. And they have different definitions and different use cases
07:18for them. And it gets really complicated. There's a lot that goes on, especially at larger companies
07:23with similar data. And so understanding how to model your data, how it all fits together,
07:28how to define it to create your metrics and your KPIs and everything, you know, further down the road,
07:33that is a really important skill to know how to do. The next one is how to use AI in
07:37your daily
07:38workflow. I am not an advocate for using AI with everything. In fact, I run several different
07:44businesses and I use AI in some of them, but not all of them and not for all the same
07:48reasons.
07:49There are specific things that I think AI does better than I do myself, just being honest. But
07:55there are also things where if I give it to AI, I know it's just going to look soulless and
07:59it's
07:59going to look bad and it's not going to sound like Alex, for example, right? It's just not going
08:03to sound like me. And so I don't delegate some of those tasks to AI. But find ways that you
08:09can
08:09incorporate it. You can use it. You can get your hands on it because there are so many free
08:12options for AI out there on free tiers and all these things where you can go and you can just
08:17try these things out for free. Now, this is not a Databricks plug. They are not sponsoring this in
08:22any way. But there's a free version of Databricks where you can use all of their AI tools for free.
08:28It's genuinely just one of the best free resources that you can use to get your hands on working
08:33with AI within a database, to create semantic layers, to learn how the business user is going to
08:38use it. It is incredible. And it is the best free resource. And so I'll leave a link down below.
08:43Again, there is no association. There's no, you know, payment here. I just really think it's an
08:48amazing resource that you guys should be using. The next one's business knowledge. I'm not going to
08:52dive into this one. Just understand how businesses work and really learn it well. That is the biggest
08:57differentiator. And for me, it was healthcare. Like I know healthcare like the back of my hand in
09:02multiple different areas. And that's where I got the biggest boost in my career was really
09:06understanding the business. The last one I want to talk about is cloud and just modern data stack
09:11literacy. And what I mean by that is just understanding how a lot of new companies and
09:15larger companies are actually working with data. When you're working in a smaller company,
09:19they could just be working within, you know, Excel. That might be all they're using. Or maybe a
09:24small CRM or something like that. And that's totally fine, right? If that's the level that you're at,
09:29if that's the level they're at, learn that well. But I would kind of push you to start learning some
09:35cloud tools and some more modern data stack. Learn how to use a data warehouse, how to use
09:40DBT, Git and GitHub. And then I recommend learning some big cloud platform tools or platforms, which
09:47are things like Databricks, Snowflake, AWS, Azure, GCP, which is Google Cloud Platform. I highly
09:53recommend learning all of these or at least getting your hands on one or two of them because it is
09:58just
09:58so important to understand how these things work. Because even at small companies, they're starting to go
10:03towards these types of tools because they're almost all all in one. You can get all you want
10:08done in one place instead of having to have four different licenses and all these things
10:12in different places. And so it is really beneficial and it helped my career immensely when I started
10:19learning Azure. And then I just went on from there to learn other tools as well. All right,
10:23so now let's get into hiring. What are hiring managers actually looking for today? Now, I will say
10:30I have a whole platform called Analyst Builder where I help people get ready for interviews,
10:34whether it's coding interviews or general technical question interviews. And we just
10:38launched our new thing, which is where you can actually do a live interview or an AI live interview
10:43to practice for interviews. So I feel like I have a really good sense of, you know, what you should
10:47be doing and what you should be looking at. And this isn't just my opinion fully because I haven't
10:52been a hiring manager for like five years, but I have friends right now who I've been talking to for
10:57the
10:57past year or two about how their hiring has changed, whether it's at small startups or it's
11:03at big tech companies. And what almost all of them said, and this is, you know, all in the blog
11:07if you
11:08want to read it more in depth, is they're like, we're looking at more thought process now. We're
11:12not doing as much coding, although they still do coding, but it used to be like 95% of the
11:19time
11:19they're doing SQL coding tests to make sure you can write SQL. Now, at least for the hiring managers
11:23that I talked to, you know, people who I actually know, they're saying we're still doing it, but it's
11:27not as important that it is the thought process. So whether you can write a join is not as important
11:32as, hey, walk me through why you need to make the join, what your thought process is on how you're
11:38going to do it, right? So it's just a different mindset. And so you still need to know how to
11:42code.
11:42You just need to know the why behind it and the how as well. They also said that domain knowledge
11:47is more important than tool knowledge. I already talked about that, so I'm not going to dive into
11:51that, but still an important thing to just be thinking about. If you're just a person who only
11:55knows SQL and Excel, then you just know the tools. You don't know the business behind how companies are
12:02actually implementing those tools. We also talked about, and I think this is a really interesting
12:06one, is that they're looking for people who can own processes now instead of being that guy, right?
12:11You're the dashboard guy. You're the SQL guy. You're this guy. It's people who own entire processes.
12:16And so they want someone who is not only going to live within a tool, but live within the life
12:22cycle of that data. So from ingestion all the way into analyzing that data to getting it to wherever
12:28it needs to go. That doesn't mean that you're going to have to do like data engineering work or data
12:32science work. It kind of sounds like almost project manager ask. You're going to own that project from
12:37start to end instead of having, you know, your typical person who's going to own each slice of the pie.
12:43Now let's talk about how the modern data stack is changing. And I will say some of the things
12:46are about to talk about are already things we've just discussed, but I'm going to put it a little
12:51bit more concretely and add a few more things. The first one is that we're looking at fewer tools
12:56and that's happening now, as well as I think is going to continue into the future. I don't think
13:01that you're going to need to know as many tools in the future. That's just my opinion. I think a
13:06lot
13:06of these platforms are becoming these all-in-one platforms where you just can do everything in one
13:11place. That was kind of the idea behind Microsoft Fabric. I don't think that ever really materialized,
13:17but it is happening kind of across the board where now instead of six or seven tools, you're using
13:22one or two. And so again, that tool mindset where you have to know all the tools isn't as important.
13:29It's really understanding how to use, you know, what you have and the mindset and the business
13:33behind it. I know I've used them a lot in this video, but I'm just trying to stay consistent because
13:37I could pull a lot of different platforms out here. But Databricks is kind of your all-in-one
13:42place these days, right? You can ingest data, you can clean data, you can create automated processes,
13:47you can get business users to look at it, you can build dashboards, you can kind of do everything
13:52in one place. Now there's lots of platforms that can do that to some extent. Databricks is one of the
13:56better implementations that I've seen, but you'll see people start to shift over to these platforms
14:01because it's just easier to maintain, it's easier to have all in one place. And within these platforms,
14:08they're writing SQL, they're writing Python, you have to know AI, you have to understand the business,
14:12you have to understand data modeling and semantic layers. It's just all in one place. So it's a lot of
14:18knowledge still that you need to know, but you don't have to go and learn all these different tools.
14:23The next two are probably some of my most controversial takes, and that people gave feedback on,
14:29and they were like, well, I kind of agree, but I don't fully agree with this. And the first one
14:33is
14:33that AI is giving business users self-analytics. And I put self-analytics in quote, and I did it in
14:38the blog, because self-analytics is a very interesting term. It's been around for many years where you're
14:44giving business users the ability to analyze data themselves. It's an incredible idea. And I actually
14:50somewhat tried to implement this back at one of my previous jobs, and it just never fully materialized.
14:56And I still don't think it is going to, but it is empowering business users to get their hands on
15:01data like they've never done before. There are a lot of tools and a lot of platforms right now where
15:05people can use AI to ask questions to their data. I'm going back to Databricks again. Let's look at
15:11Databricks 1. This is a prime example of what I think is kind of the best implementation of this,
15:16where data analysts are building out all these semantic layers, all these business rules and
15:20definitions, and doing all the heavy lifting behind the scenes, so that a manager or somebody on a team
15:26who isn't as, you know, technically literate, who can't go and query their own data, can just ask
15:31questions about their data. And it's going to give insight. It's going to give numbers. It's going to
15:36give all sorts of visualizations for this data to that business user. Now, what I've found is that
15:42the more that businesses use this tool, the more work it actually creates for the data team. Why is this
15:48number like this? Why does it not match from last quarter? I ran this today, and then I ran this
15:53yesterday, and the numbers are different ever so slightly, and I don't understand why. And so this
15:59is where the data team is just going to be slammed with all types of ad hoc requests. I'm just
16:04telling
16:04you. It's going to happen. But this is going to be something that is going to continue down the road
16:09where companies want to empower their users to use the data. They want to make use of all this data
16:15that they've collected. And so this is a great way to do it. It just is not something that's going
16:20to be
16:20self-contained. It's not something that, hey, the business user can just ask their questions, and
16:24they'll be done. It's just going to lead to more questions. It always has. I just think it always
16:29will. My last take, and this is the one I actually got messages about, and more people disagreed with
16:34this than anything else, and I still believe this. But I think we'll slowly move away from dedicated
16:40BI tools like Power BI and Tableau. I'm going to use a different example for this one because I just
16:46recently built something with this for one of my clients, which is within BigQuery, if you go into
16:51BigQuery, you can use Data Canvas. And Data Canvas, you can build all sorts of visualizations, and you
16:57can have AI help you with these visualizations. Now, the reason why I think these are going to be
17:04really useful and kind of slowly chip away at the dominance of Tableau and Power BI is because a lot
17:10of what Tableau and Power BI is great for is it's really great for when you're connecting with
17:15clients, business users, different departments. But a lot of people don't want to pay a separate
17:21license or these extra fees just when it's like internal for your team, which you currently have
17:26to do, when it's already built in to a lot of these platforms. So BigQuery, I can just go into
17:33Data
17:33Canvas. I can share that with people. They can look at the queries. They can see the visualizations
17:37all in one place without having to have those other licenses. And so from purely just like a
17:42money standpoint and a practicality standpoint, I really think that even with Tableau and Power BI
17:49getting a lot of AI integrations, I don't think people are going to want to use those for their
17:55everyday use cases when they can just build them a lot quicker within this tool that they're already
17:59working with the data in instead of trying to bring the data into Tableau and Power BI and connect it
18:05and, you know, do all these things. And it's just a longer process. That's just my opinion. So I think
18:09that that is going to happen, although many people are like, no, you know, Tableau and Power BI are going
18:14nowhere. And we'll see. I don't think they're going away. I never said they were going away. I just think
18:19that they'll slowly lose their dominance. I think that these integrations with dashboards within the
18:25actual systems where you're working with your data are just going to kind of chip away at that over
18:29time. In the end, I know that, you know, with AI comes change, but it also can be really scary.
18:37And the job market is just horrible right now. And I understand that. I actually don't think
18:42that's as much AI as maybe the media would say. It's just a bad economy, especially in the United
18:49States. It's just not a great economy to be finding a job right now. That's just the fact. I think
18:54these
18:55types of things ebb and flow, right? They go up and they go down. There's huge hirings. There's huge
18:59firings. And over time, we're having this big downswing. But I think over time, we will see a
19:04big uptrend again. So I hope that walking through this, you can kind of look at yourself and say,
19:09where are my gaps? What are things that I need to be working on? And you can take action on
19:13that.
19:13Because one of the worst things that you can do right now, especially in this, you know, era of AI
19:18is
19:18just to not progress, right? And I hate to say that you need to keep learning and keep progressing and
19:24you
19:24need to do this and that. But it's true. Like we all need to adapt. We all need to be
19:30learning,
19:30you know, the things that are important today. If you're still only working with Excel from 20 years
19:35ago, you've missed a lot of progress in the past 20 years or so. And so it's really important to
19:40keep up with things. We are in a technological age and we work in an industry of technology.
19:45So we have to keep up with those things. It can be scary. It can feel overwhelming. But luckily,
19:51I have created so much free content on all the things that we're discussing in this video.
19:56I have tons of tutorials on Databricks, AWS, Azure, AI, SQL, all for free on YouTube. And I will leave
20:03links down in the description for that. I also have paid stuff where you can go even more in depth
20:08and
20:08build really advanced projects within all these things. And I have more that's not on YouTube.
20:13And that's all on Analyst Builder. And so go on to Analyst Builder and check those things out.
20:17And these are the types of things that you can do to not just stay ahead, but also to set
20:22yourself up
20:22to be competitive in this job market instead of the job market of five or 10 years ago. That is
20:27my
20:27opinion. And so I do want to hear your thoughts. I do want to hear your opinions down in the
20:31comments
20:32because I promise you I'll read them. And it may help other people see a different viewpoint as well
20:37because again, I don't think my viewpoint is perfect, but I think is pretty informed to the best of
20:44my ability. So with that being said, I hope you enjoyed this video. If you did be sure to like
20:49and subscribe and I will see you in the next video.

Recommended