What I’ve been up to

I haven’t posted in a while, so I thought I’d let you know what’s going on and what I’ve been doing in the past two years.

I started working fulltime. Yeah, that’s about it. I really struggle to find the motivation to sit down on a free day and work on projects that I can show here for several reasons.

First, I feel like the things I can achieve in a few hours outside my regular working ours are nothing compared to what I do in my job.

Second, I can not show projects here that are too similar to the things I do at work.

But most importantly, most of the time I rather spend my free time doing things that seriously relax me, like going to a spa, and spend my time getting the physical exercise I feel I need for my health and wellbeing.

Anyway, I promised myself to contribute a little bit more to this blog, at least in wintertime.

Let me tell you a bit about my job: At my job, I’ve had various different projects that involved data analysis for clients in the FMCG industry.

I started there as an intern. I worked on NLP projects analyzing short comments from social media and used topic modeling algorithms to attribute topics to short text comments, and sentiment analysis to classify the comments in positive / negative / neutral sentiment.

I also did some trials with the google vision API to classify images from social media into different categories.

Then, there was my thesis. I administered a survey questionnaire to German shoppers to capture their response to novel food products and then used Bayesian Methods and some fancy regression methods to crystallize the most important factors influencing food product success.

When I started working full time, my projects kind of changed.

Now, things had to be done for the clients. Data had to be analyzed, expectations met and stories to be told. Our analyses had to be “sold”, actually. I started with some surveys about products, made some customer segmentations, and ended up in projects involving scanner data from large stores.

In these store scanner data projects, I got involved in the world of causal inference: Using difference-in-difference methods, and by creating synthetic control groups, the actual effects of changes in store layout or marketing actions could be quantified.

Since I can not share the actual code with you, I made some kaggle dataset regressions using rstanarm, which you can view here.

Other projects involved the creation of web-apps, which involved programming the backend, frontand and deployment. The learning curve was very steep, especially when my manager, who created the app and the analysis code for the scanner data projects, left our company.

By now, things have gotten calmer and I work on my projects with more routine and self-confidence. We introduced some new frameworks (gitlab, azure databricks…) and continue to improve our teamwork and codebase.

Even if the work as a consultant can be stressful at times with ambiguous demands, changing priorities, messy data and short deadlines, I feel like I’m improving my knowledge and skills.

I hope to find the time to take a breath and reflect like I’m doing now from time to time. Cheers.