Python is Changing – Here’s What’s Coming

Key Takeaways

Just like programming is evolving, so is Python. The way that we use Python today is drastically different than five or 10 years ago. And in this video, I want to get into that in a lot more detail. Now, I actually had a look at the Python developers survey. This surveyed over 30,000 developers specifically focused on Python. And in this video, I'm going to break down some of the most interesting results. Now, this survey was conducted by the Python Software Foundation as well as the Jet Brains PyCharm team. So, what I'm going to do is dive into all of the results. I'm going to share with you. And by the way, if you want to check it out, I'll leave a link to it in the description because I'm just going to go over some of the most important parts and what it actually means for you as a Python developer in 2025 and beyond. So, I've got the survey open. You can access the same page from the link in the description and I'm just going to skip through and cover what I think are the most interesting insights and what that actually means as someone who's a Python developer. Now, one thing to note is that this was done in November 2024. Hence why we don't have the 2025 version yet because it's going to happen in 2025. The results were just kind of aggregated recently. And then this was also surveying 30,000 Python developers. So people that mostly write in Python as you're going to see as we go through this survey. It wasn't surveying just all random developers. So keep that in mind. This is very biased towards Python developers as it should be considering it's the Python developer survey results. Anyways, you can also see the previous uh years if you want and you can kind of see the trends that are going on there. Okay, so let's keep going here. First thing, general Python usage. So obviously if we're surveying Python developers, well, most of them are going to be using Python as their primary language. You can see that's 86%. And then what's interesting to note is that along with Python, most people are using JavaScript. You can see that's 40% as a secondary language. And then things like SQL, HTML, CSS, and that makes a lot of sense considering what Python is used for, mostly web development and data science. Cool. Let's keep going. Uh, another interesting point here is that most people that are writing code with Python professionally have been doing this for less than one year or in this case less than two years. But 50% of all Python developers, professional Python developers have no more than 2 years of experience. So if you are someone who actually does know a lot of Python, you become an expert in that field. You have 5 6 7 8 10 years of experience in Python. You're going to be quite rare as you can see here because most people are kind of beginners when it comes to professional experience. And then in terms of how long have people been programming in Python, what this tells me here is that most people have learned Python at some point in their life. They've been doing it for a while, but they just haven't done it professionally. Hence why we see the largest um kind of category here, 30% of people having 3 to 5 years experience. That doesn't mean professional experience. It just means they've been programming in Python for that long. And then that leads us to this, which I think a lot of us probably already know, but it's worth noting that Python is the most popular programming language for learning to code. Hence why I do so many tutorials on this channel and I typically recommend Python as the first language to get into. Okay, so let's keep going here. There's a bunch of other stuff that we can look at, but again, not all of that is interesting. What I do want to get into is what Python is actually used for. People ask me all the time, why should I learn Python? What do you actually use it for? Well, we can see here based on the survey results what the most popular uses are. Now, of course, we're going to have data analysis and web development up there. And then we have machine learning, data engineering, web scraping and parsing. That's something that's getting a lot of attention recently. Academic research, that's not as interesting to me. And then we can go through the rest of them where it gets, you know, quite small down here for things like MLOps, game development, embedded development, etc. So, if you go here, you can see kind of the main categories. Again, we can skim through those. I don't think it's too useful. We pretty much all know that Python is used mostly for web development and the machine learning, AI, and data analysis. And if you go down here, they kind of rank their primary, secondary, and hobby activities. And we can see when it comes to hobby activities, a lot of people are doing uh it's kind of hard to check this here. Programming of web parsers, scrapers, and crawlers, which is kind of interesting. That's, you know, a pretty big jump there uh compared to the other sections. Okay. Now, this is where I think it gets a lot more interesting in terms of the frameworks that people are actually using for these different activities. So if you're someone who wants to become a Python developer, this your favorite language, you're writing it all the time, you should probably pay attention here to the frameworks people are actually using in professional development because that's probably what you want to learn. So we can also look at the trends and the growth. And I mean this actually even relates closely to me. I used to always just use Django and Flask. Now I pretty much just always use fast API for doing my APIs. And we can see most people are doing that as well. So of course we have fast API, Django, Flask and requests, async.io as well. But most people when it comes to building APIs now are going with fast API. So that's definitely something that you'd want to focus on if you wanted to become a professional Python developer. And you can see all the growth there. I expect that's going to continue to grow as more and more people learn that framework. Again, few things over here in terms of data science web development, but generally speaking, Flask, Django, fast API request is a pretty easy one to learn. So of course that's going to be up there, but most people favoring fast API for API development and I would assume Django for more full stack sites. If we go here, we can see a few other web framework crossage. So a lot of these frameworks are kind of intertwined to one another. So for example, Starlet is used all the time. AIO HTTP like they're dependencies of each other, which is kind of what's happening right here. And we can see Django, you know, is used with Django West REST framework 93% of the time, right? Fast API is used with HTTPX or Starlet, of course, right, for running the server 92% of the time. And you can kind of go through those, but I think, you know, that's pretty straightforward. Cool. Continuing here, other frameworks and libraries. Beautiful soup. So if you're doing any kind of web scraping or parsing, that's a really interesting library to look at. Pillow, this is for image manipulation. Um kind of loading in dealing with images and computer vision related stuff. Pyantic, that's going to be super popular for uh frameworks like fast API where you need to do the type validation. Open CV again for computer vision. I'm actually surprised that we have these tokin, piqt, uh scrappy, and pygame. Obviously, Pygame being my favorite Python module, but Tequiner and PIQD is interesting because you can use those to build uh what do you call it? Graphical user interfaces or desktop applications. Actually interesting to me here that that's increased a little bit from 2023 to 2024. Unit testing framework pretty much everyone is just using piest or unit test which is built into Python. I actually have a tutorial on this. Definitely recommend you learn it. It is quite simple. And then of course all of our great Python developers, 36% of them not testing at all. So the next interesting part of this survey to me is the data science section. Now first of all 51% of devs are doing something related to data science which obviously makes a lot of sense especially with AI becoming super popular now. We need data to train those AI models. And when it comes to the libraries that people are using of course we have pandas and numpy. I'm sure you guys probably know this but if you want to get into data science you need to know these modules. And then we have a few other ones that are a bit less popular like polars airflow inhouse solution dask etc. But pretty much everyone is just using the go-to triedand-true pandas and numpy. I'm going to skip the data versioning here. This one's interesting as well. Streamlit is the most used by far now for uh data dashboards. And then we have plotly as well. This is what I use all the time now. I recently found out about Streamlit maybe a year ago, year and a half ago. Since then, it's been my go-to. If you haven't used that before, definitely recommend it. It's super easy to make basic web interfaces with Python without having to know any JavaScript, HTML, etc. in like a few lines of code. You can make, you know, really interesting dashboards and display all of your data. Going to skip over a few of this other stuff, but let's get into the ML model training and prediction. So, interestingly, most people are still using Scikitlearn. I think that goes to show that while we have all these LLMs and crazy models and stuff, uh, people still just need some more basic AI solutions or machine learning solutions and scikitlearn is pretty much perfect for that. And then of course we have PyTorch and TensorFlow. Interesting to me that PyTorch is more used now. Uh, previously a few years ago, I remember TensorFlow was more popular, at least that's what I was using all the time. And then we have a few other ones like Scypi, KAS, etc. which is kind of just an extension. hugging face transformers and you can see that's got a big boost recently 6%. I think a lot of people and I've seen this as well are going towards hugging face for a lot of those models and then of course we have the cross usage here. So for example natural language to toolkit is used all the time with scikitlearn and we pretty much use everything in combination with scikitlearn to get these tasks done. Same thing here with pietorch. You're going to use a lot of these in combination with another uh scroll through the rest of this. We can see that most people are using Jupyter notebooks when they're actually doing the training and kind of data exploration, which is interesting. And I'm going to scroll through to the next section. So now we're getting into the development tools. Now, this is pretty interesting. Obviously, this doesn't add up to 100%. I assume that's because people are using multiple operating systems. Most people are using Linux, then followed by Windows, believe it or not. I know a lot of you guys think people only develop on Mac and Linux, but a lot of people write code on Windows. And then we have Mac OS, BSD, and other. And what AI tools are most people using? Chat GPT. That's interesting to me. I would have thought more people would be using things like C-Pilot or built-in idees or AI idees. However, I guess it's just first of the market. Chat GPT is the one that most people are comfortable using. Uh especially I guess with more beginners in the Python ecosystem. GitHub Copilot, Google Gemini, Claude, Visual Studio Intellode, Code GBT plugin, etc., etc. Tab 9. So we can see the more traditional autocompletes which used to be popular a few years ago are kind of taking more of a back seat now to the full LLM models which is an interesting insight for me. OM. So if you know OMS, object relational mapping, most people just using SQL alchemy, we can see for web development and data science and that's becoming more and more popular, which makes sense because in my opinion, it is just the best. And then Django OM, of course, if you're working in Django, you're going to use their OM. Databases, another interesting one. Postgress SQL getting some growth here. SQL light, same thing. MongoDB, still not super popular. So most people are opting for your traditional SQL tablebased databases uh with Postgress, SQLite and MySQL at the top and then some of the more NoSQL or caching databases a little bit lower down. Then we get to CI and CD. Not surprisingly to me, we have GitHub actions followed by GitLab CI up here. Makes sense. Just the simplest solution by far for doing continuous testing and integration. So I can see why those are the most popular amongst amongst sorry all of these Python developers. Most people are not using anything for configuration management, documentation, markdown, of course, that makes sense. Swagger, Sphinx, Postman, etc. And how do you typically work in a single Python file? I open the entire project that contains the file in the IDE. Makes a lot of sense. Some people using a command line editor. I can't imagine why you would want to do that unless you're on a VPS or something. I just open the one file in the IDE. Okay, that's interesting to me. I think these are probably more of the professional developers that are answering this 58%. Then main editor and IDE. You guys might find this interesting. We have pretty much just Visual Studio Code and PyCharm close behind. Makes a lot of sense to me. Those are just the most popular with 3% of people using Vim. Leave a comment down below if you are a Vim user. I'd be interested to hear that. And just a few last interesting points to go over here. Python packaging. So what are people using? See most people still stacking with Venv or Venv or however you want to say it followed by virtual env. Then poetry pip env. And we can see UV obviously a very new tool already grew to 11% which is very interesting and this is what I'm using pretty much all the time now. So I would definitely predict that next year UV is going to have a huge growth. I don't think it's quite going to be as popular as Venv. But I would suggest if you haven't already checked out check out UV. Very popular, very fast, very easy to use package manager and it's been my go-to in Python for the past probably eight months since I discovered it. Let's keep going through here. Anything else interesting? what tools you use to manage dependencies. Same thing, UV showing up here. And then, of course, we have pip. I mean, pip is just going to be used everywhere. And then lastly, of course, requirements.txt, pipro.toml, setup.pies makes a lot of sense. Now, another thing worth noting here is that Rust is picking up some popularity when it comes to building binary modules for Python. So, in the past, it's mostly C++ and C, but we can see Rust is slowly growing. And I believe we're going to continue to see that trend as more and more people learn about Rust and understand kind of the benefits and the usage. If we scroll down, there's a bunch of other stats. We can see, you know, 89% of people are male. I think that's probably close to just the programmer demographic in general. And then for the age, we can see most of them are sitting in between 21 and 39. Makes sense. Again, that's just the core demographic of programmers. Most programmers do not last that long in their career and by the time they're in late 30s, early 40s, as we can see here, they start to drop off pretty quickly. Be curious to see this compared to some other career paths, but I know in programming specifically, you typically don't find crazy old programmers. By that time, usually they're rich enough and they just retire. At least that's my experience. Anyways, that's kind of the main stats here from this. Now, look, there's a bunch of stuff that I didn't go over. I just wanted to cover what I personally thought was interesting and that I think you guys should be aware of. Generally speaking, what I'm seeing here with Python is that while it is evolving, more and more people are moving into data science using AI. We saw almost 51% of people are doing some kind of data science related work. People mostly are sticking with the tried andrude modules, the ones that just work that have been used forever. And then there's a few new tools that people are using because clearly they are just better. For example, we have UV, we have Streaml, we have everyone really pushing more towards fast API using paidantic with that. So, I think those are some interesting trends and if you're someone who's still using some of these older modules or you haven't switched over to a different package manager or you're still writing everything in Flask for example, then definitely consider kind of hopping on the trend and going with what most people are using because that's what's going to be used for most of the jobs, right? Or for most of these companies that are going to be hiring you. Of course, there's lots of other insights that we could dive into here. We could spend hours talking about this survey, but I'm curious to hear what you guys think. If you go through the survey or you just watched this video, leave a comment down below. Let me know what you think of the future of Python in 2025 and beyond. If any of these stats surprised you, if any of them, you know, you thought were quite normal, there's any other insights I didn't share that you think are interesting, I would love to hear them. With that said, thank you very much for watching and I look forward to seeing you in another video. [Music]