The week I found my (Computational Social Science — CSS) Tribe 14 April -20 April 2023
My first CS conference acceptance, Fang’s sophisticated code and a twitter thread on coming to CS from outside the field
What happened this week?
I got accepted in to the 9th International Conference on Computational Social Science (IC2S2 short) for a poster presentation on the paper I wrote over the winter, focusing on leadership in digital transformation using topic modelling. Besides the massive sense of achievement as a newcomer to the field, when I read the proceedings from previous years I had this sense of: this *is* my tribe.
This is an important realisation — I am looking at applied computer science, (more specifically topic modeling techniques) into other domains, such as business studies and medicine. That’s how I am aiming to push the boundaries of text processing pipelines and experiment with various tools and techniques.
Here’s a sneak peek at my poster… so proud. Thanks Canva for a great medium to my creativity.
What was hard?
I met with my colleague Fang who developed a technique called Query Driven Topic Modelling (QDTM for short — which can be found here), who kindly shared code adapted to my problem. The data that I am using is set in PDF format, and the topics that I want to explore are well-hidden within the data, so an initial, automated key-word search for the relevant passages of text is helpful to prime the model. My initial approach had been entirely manual which doesn’t suit for scaling up as I plan to do in the next stage of my research.
Despite the code being ready to go (at least in Fang’s machine), it took us two hours to get it running in mine. The sense of frustration started when I could not even run the first line of code. I must say I was in some ways relieved that Fang was puzzled himself in certain areas and had to look up solutions to the error messages we found.
I realised:
This *is* what means to code. I tend to think it’s only pressing the button and things running that means the process has been successful. I need to appreciate the whole process, including writing the code, making it legible and clear, debugging and running it.
I was grateful for Fang being in my tribe. While I understand and can code, the level of sophistication Fang can achieve is something that will take me years to build. I bowed silently in humility while reminding myself that coding is a practice.
What did I enjoy?
There was a moment of jubilation when the code ran…
Just to find out that my run of the model gave different answers to his.
We parked it for the week but we’ll revisit the results. I will be reviewing the code in detail this coming week. Totally felt ‘coded out’ after that session!
What did I code/write/reflect on
Code Prompted by Fang, I coded for the first time a virtual environment to run python out of my visual studio shell. Turns out I was running it from the base which is not good practice (!). More on virtual environments here.
I wonder if that’s what was needed to be able to run things that were not running before because various versions of other packages….LDAvis here I go!
Write I progressed writing up about digital transformation in healthcare and the vision of ‘Health 4.0’ (might write a blog on this on my other publication, ‘Digitalling the NHS’).
Reflect on The proceedings from previous IC2S2 conferences are very interesting and can be found here. This has also opened my interest in reading conference proceedings more in general to quickly get a sense of where the field is.
I wrote a twitter thread on some Computational Social Science articles I found. This will be the beginning of a section in my dissertation, and potentially a follow up blog.
