API and ML

In the lecture we talked about the API of a few different servers. It reminded me some of the projects that I have done and needed to gather some data. My projects have been mainly designing or improving AI models. The first step in ML is gathering data. For one of my projects, I needed to download all the available bacterial genomes, and NIH has an API that we could use to filter and download the data that we need. I was working on this project when I did not have any experience using APIs and programming. However, documentation and the paths were designed in an intuitive way so that I caught up on working with the API easily. Another time, I was working on an NLP project and needed to gather data from Reddit, but I had some challenges simply because the Reddit data is not structred in a scientific way that genomes data are structured. However, Reddit has a thorough documentaion.

After this lecture, I learned that the API design can affect the efficiency of data gathering for ML. This is something that I had never thought about it.