Last week I attended a short 3-day workshop on “Transitioning from Academia: Developing a Career with Python” with an emphasis on data science, which was organized by PIER and faciltated by Alex Britz.

The scope of the workshop was quite broad, jumping from strategies for the job search like networking and preparing a good landing page, to a whirlwind tour of some technical aspects of doing the job of a data analyst/scientist, like standard packages (pandas, numpy, seaborn etc.) and newer tools like ChatGPT and Github Copilot, as well as methods to share and present results, e.g. data visualizations and Streamlit.

One of my key takeaways is that in a non-academic setting, ultimately people only want to know two things: what problem we are solving, and what the results are. This was not news to me, but I think the workshop really hammered it home, particularly when Alex blitzed through a project in 30 minutes with the help of ChatGPT. Actually this is not too dissimilar to communication in academia, especially when speaking to someone in a different field of research, though some people, perhaps out of politeness, may follow up with questions about methods employed or challenges faced. One may even argue that a well-written academic paper should have a similar goal-oriented emphasis, in particular in the introduction, where the main theorem is clearly stated. Regardless, I should practice being to the point when describing the goals and outcome of my work.

Reflecting on my experience doing data analysis at my previous job at Albert Einstein College of Medicine attempting to predict the cell-to-cell contact based on the gene expression in each cell, I was constantly experimenting with the data, and despite producing many Python scripts and using many different models, I never felt that I had arrived at a conclusion, always uneasy about the outcome of my experiments. While certainly my inexperience in data analysis contributed to this feeling of unease, I also think that the environment, being that of a biology lab focused on fundamental research, suggested that all of the results I obtained should only be taken as suggestions for further research with real experiments, rather than anything definitive or worth writing about. (I should note that I was mainly employed as a web developer.)

Another thing I appreciate about the workshop was being in the company of academics who are leaving academia, many of whom also do not speak German very fluently. They shared many of the same concerns I had, e.g. being “overqualified” with a PhD, competing with younger data science majors, falling short of a 100% fit with the job description etc.

Finally, I enjoyed collaborating on a data analysis project. We found a data set on Kaggle concerning the properties of minerals, which interested us both. We wrote our own code, but it was still really helpful to be able to bounce ideas off one another, especially since neither of us knew much about minerals. Here is my notebook.