Behind the Scenes at Gapminder: A Story about Data and Visualizations

on YouTube ↗43m

" summary

Pipeline shows concrete work: large-scale surveys get open numbers harmonized with custom Python ETL recipes and Apache Airflow, transformed into a DDF tidy multidimensional model, validated and published to GitHub and a public API, then served from a struggling statistical backend now using CSVs on S3 as a stopgap to the waffle server. Visualization stack is JavaScript/TypeScript with animation-first viz RB and an Electron offline tool; claims include WHO vaccination at ~85% and long-term population ~11 billion, with explicit failures and tradeoffs described.

AI-written summary · about

" chapters (9)

1. 0:00Start
2. 8:59STATS SA
3. 21:02Multidimensional statistics
4. 22:20DEVELOPMENT GOALS
5. 23:51Tidy data
6. 25:59Load Validate resulting datasets
7. 29:14Visualization Grammars
8. 40:2510 Dramatic instincts
9. 40:54Factfulness rules of thumb

" related talks

• 35mVisualizing research data: Challenges of combining different datasourcesJuha Suomalainen · PyData Berlin 2016
• 22mGovernment dashboards and open data.Paul Haesler · PyCon Australia 2015
• 36mData-visualisation with Python and Javascript: crafting a data-viz toolchain for the webKyran Dale · EuroPython 2015
• 58mTrustworthy Data VisualizationKieran Healy · posit::conf(2025)
• 47mData, Design, CodeShirley Wu · beyond tellerrand Düsseldorf 2019
• 49mSaving the world with JavaScript: A Data Visualization storyMarko Letic · JS Fest 2019 Autumn
• 33mData Science & Data Visualization in PythonRadovan Kavicky · PyData Berlin 2017
• 03h 07mData of Unusual Size: Interactive VisualizationPavithra Eswaramoorthy, Dharhas Pothina · PyCon US 2023