I'm mostly doing quantitative analysis and data visualisations. For this, I use lots of:
- Python (pandas, numpy, scikit-learn, jupyter, spacy) and node.js (puppeteer) for scraping and data analysis (statistics, NLP, machine learning)
- Typescript, React, Svelte, d3, SVG, canvas, WebGL, topojson, RxJS or firebase for data visualisations
But a part of my job is also to industrialize and build tools to improve our processes, reuse components through dedicated libraries or scaffolders.
- design, development
- a live dashboard for the french local elections
- design, implementation (including a routine to retrieve and process the data)
- a search engine to analyse political alliances
- processing, design, implementation
- processing + design + implementation
- for this piece, I've analyzed thousands of images from fashion collections with a tensorflow model to extract silhouettes. Then I've used a quantization algorithm to get color palettes for each brand. Finally, I build a scrollytelling visualization with arcs patterns to show the generated palettes.
- a tool to visualize the evolution of Ligue 1's ranking, made in d3
-
a scrollytelling piece and a Sankey chart variation, made in d3
-
Le confinement a-t-il permis d’endiguer le nombre des décès ?
-
Votre ville est-elle un bastion de la droite ou de la gauche ?