Miscelaneous projects

I developed a set of interactive tools to track the development of the COVID-19 situation in Argentina. They feed on microdata published by the Ministry of Health on a daily basis, and aim at bringing these large datasets closer to the general public, facilitating epidemiological surveillance by both experts and non-experts.

The first product tracks COVID-19 cases in the whole country at multiple levels (e.g., provinces, cities and demographic groups), and was developed with the advice of a group of Argentinean epidemiologists and physicians. It also combines cases data with districts' socioeconomic indicators to map inequality in the spread of COVID-19 across regions. It was deployed using Google Data Studio, although Python and SQL run behind stage to parse and manipulate data, and to create the variables. It can be accessed following this link.

The second one focuses on the city of Córdoba, Argentina's second largest and my place of birth. Leading analysts and media follow the situation of Buenos Aires, and there is little information about the development in the rest of the country. This tool was developed entirely in Python, using Bokeh for the interactive plots. It was deployed with Heroku and can be accessed following this link.

The third product, deployed around the end of March 2021, uses anonymized microdata on COVID-19 vaccine doses applied in the whole country. It was developed entirely with Python and Bokeh, and deployed with Heroku. It is available here.

I am a huge classical music fan and I like to explore music in breadth and depth—discovering both new works and different versions of known ones. Listening to music many hours a day, I often find it hard to decide what to listen. Music suggestion algorithms in Spotify, Idagio and Youtube do not help, because they were not built to deal with the complexities and subtleties of classical music. This project addresses this issue, exploring a new way for measuring the distance between classical composers and their works. The goal is to help creating playlists with suggestions based on music (dis)similarity, combining machine learning and techniques common in patent analytics and bibliometrics.

Click here to read more about it in a very preliminary draft.

Are you also interested in music or in this project? Contact me at manuel.gigena@kuleuven.be!