Building CSV-powered tools for social sciences

CSV is ubiquitous in social sciences and in the humanities. CSV data is indeed the perfect bridge between social scientists, accustomed to dealing with tabular data, and research engineers needing to process the same data. That is why SciencesPo's médialab has been building many of its Open-Source tools around CSV files, from well-designed web apps such as Table2Net to convert tabular data into graph data, down to powerful CLI tools such as minet to collect data from the web or xan to process tabular data using constrained resources. What's more, the CSV format is aligned with an ethos of sobriety and does not require overpowered hardware to be processed. This is very important to us because our public is mostly comprised of researchers, students, data journalists and other members of civil society that do not have access to powerful machines & servers. This talk is therefore an occasion to tell the tale of 10 years of building social sciences tools around CSV data and to be a testament to our lab's love for the format.