Seeking Suggestions for a Julia Tech Talk to Python Data Scientists

Hello everyone,

I’m preparing to give a tech talk to my colleagues, who are primarily data scientists and developers using Python. Our products are software solutions related to the mining industry. The focus of the talk is on introducing Julia as a potential tool for our projects. I plan to discuss Julia’s performance benefits, package management, and notebook frameworks. However, I’m looking for additional suggestions on topics that would be particularly engaging for a Python-centric audience.

Our team has recently encountered performance limitations in a new project, and after benchmarking our prototype code in Julia, we’ve seen promising results. This tech talk is also a strategic opportunity to generate interest in Julia among our team members.

What other aspects of Julia would you recommend highlighting to showcase its strengths and advantages over Python, especially in a data science context?

Thanks in advance for your insights!

Python interoperability, so they know they don’t lose access to all their old tooling in a pinch.


+1 for interopability. PythonCall.jl (and CondaPkg.jl) is great for that.

Multiple dispatch also. One thing that bothers me as a Julia & Matlab coder who sometimes ventures into Python is that functions are like…properties of objects? Instead, in Julia, functions are “their own thing” and if you want to define some already existing function on your new object (type) you can do that.


Maybe you could also present some code snippets of Python and Julia side-by-side (of some simple representative workflow maybe) to show that Julia does not look too different from Python, i.e. not littered with type annotations, weird brackets, semicolons… Many of the basic language feature are similar to python. And then show some timings of the snippets :slight_smile: Maybe it would be interesting to talk about compilation time at that point as well. I don’t know how your workflows typically look, but if there are lots of small Python scripts worh short runtime then you will need to change this style when transitioning to Julia towards bigger/longer running scripts.

But then again showing code in presentations needs to be done quite considerate to not be overwhelming.


I’ve just done a little hands-on intro to Julia session at the International Microsimulation Association conference. It went fairly well I think, even though I’m not much of a programmer by the standards on this board. My suggestion: don’t worry too much about discussing masses of features; instead, walk your colleagues through making something simple but fun. We built a little toy tax-benefit model using just Pluto, Plots, and some live data.


Nice idea. This should help illustrate the ease of use and the performance enhancements offered by Julia. I remember coming across a Python-Julia cheatsheet that could be very helpful for this section.

One thing I want to be mindful of is my initial misunderstanding when I started with Julia two years ago. I had thought it was very similar to Python, which wasn’t entirely accurate. It’s important for me to ensure that my presentation doesn’t give the same misleading impression, while still highlighting the strengths and unique features of Julia.

This is great. The presentation will be in Pluto, but I should also consider creating something minimal and fun during the presentation itself.