It’s the last month of 2020, and what a joy the year has been. Life is now pretty same-y, and whatever I’m doing, I’m doing it from home. That makes it easy to feel like nothing is changing, so I wanted to spend a few minutes actively reflecting on what I’ve done and learnt this year. Normally it would feel more obvious because the year would be demarcated with conferences, courses, hackathons, and I’d be meeting people at all different stages of their data science or tech careers. I’m still doing a lot, but the experiences are less clearly delineated, at least partly in space. So indulge me. Here are some of the things I have learnt most about when I think back to the start of the year.
Python
When I started my current job, in November 2019, I had not spent a lot of time working in Python. Although there have still been chances to work in R, as the rest of my team use Python, I’ve naturally been using it more. A lot more. Now I’m coding in Python almost every day. Even if I still prefer R, at least I feel pretty confident with Python.
And it’s not been totally smooth! And what’s weird is that I can’t really put my finger on when things changed from being really hard to fairly natural. It didn’t just click all at once one day - and I still have plenty to refine. But I know what I’m doing, and I feel like learning has helped broaden my approaches to problem-solving, data structuring and working style.
Topic Models
I’ve worked a lot with topic models this year, and I really enjoy them, because I find language such an interesting area. I didn’t realise a year ago how many different approaches there are to topic problems, and I’m still finding more!
All this work has been in Python, so next year I might try exploring topic models in R in some posts - a few times I’ve been recommended the stm package.
R under the hood
In the summer, I joined a book group in the R for Data Science slack community, where we are reading through Advanced R by Hadley Wickham. It’s designed to give people a better understanding of R as a language and hone their programming skills. Coming to R as my first language, I didn’t have any computer science background, so it’s really valuable to get this kind of detailed instruction.
This is something else where I’d like to post more - for example, I have been preparing the session on R6, and I think writing it up in a post would help elevate my understanding of classes and object-oriented programming. So maybe look for that in January!