A live coding demonstration of how to get some basic analytics from Instagram
In this talk given at the ChiPy Data Science Special Interest Group's monthly meeting, I started by setting out some basic questions that a social media marketing department might like to answer, such as "What types of posts seem to work best for us?" and "Which posts seem to get a lot of engagement?" and "What is our average likes per post"?, then I walk through the steps for how one can process Instagram data to come up with these answers.
Jupyter Notebook for the talk is here. It was my first time giving a "technical" presentation to a large group, and I really enjoyed the challenge of coming up with a talk that would be educational for the unitiated, interesting for the experts, and engaging overall. Also, when I was first getting started in data science, I found the lack of explanation in technical presentations like this to be frustrating at times. Frequently, methods are used without any explanation, leading the attendee to often not being able to implement or extend the new knowledge after the talk is over. I tried to step through the code slowly enough to spin out the detail, while trying to keep it fresh enough to not bore the experts in the room.