5 Inspiring Computational Notebooks of January 2022
--
Welcome to the first edition of a monthly summary of what is hot on the Observable Dataviz platform. There is so much activity it is sometimes hard to stay up to date and not miss anything juicy. This newsletter is dedicated to finding the gems. Last month two notebooks tweeted by the trending bot, in particular, found some resonance in the broader community.
First, the prolific @fil scored a hit with a technical dimension reduction technique for summarizing time-series data called topological subsampling. It’s pleasing to see technical posts do so well, I think Fil did a fantastic job of explaining the technique very well, and my hypothesis why it did so well is that it ultimately solves a very practical problem of transmitting just the salient information very efficiently. You can use the technique to speed up various parts of Dataviz, including bandwidth minimization or speeding up rendering pipelines.
The second big hit was @neocartocnrs introduction to his library Bertin.js. Bertin.js is a geo library to help render beautiful thematic maps. We are very excited because @neocartocnrs is a long-time Observablehq user and professional cartographer, and his library seems to be a manifestation of the move toward declarative style DataViz that the Observablehq community is moving towards. We wish you good luck Nicolas Lambert with Berin.js!
Of course, there were many other great notebooks, check the rest of January’s trending notebooks below! Also if you have not seen, there is a list of the 100 notebooks in 2021.
Time series topological subsampling by @fil
Using topological data analysis (TDA), we can extract meaningful points from a time series. This can be used to simplify a line (to take it from millions of points to a few hundreds), or to add some labels to a chart, highlighting “important” points such as relative extrema. The approach is inspired by (my reading of) Paul Rosen