Key takeaways
- Why H2O chose air pollution forecasting in India as the hackathon theme, using real-time data from the Central Pollution Control Board
- How the competition ran end to end on the in-house H2O Olympics platform, built on H2O Wave, with a live leaderboard for 250-plus registered teams
- The judging criteria beyond leaderboard score: solution completeness, storytelling, and a bonus point for using H2O products
- What the top teams built, including full H2O Wave dashboards with data exploration, model comparisons, and station-level distribution plots
- Why leaderboard rank alone didn't decide the winner: the eventual first and second place teams weren't the top scorers going in
- A shoutout to customer team CBA, which placed in the top 10 alongside data science teams from across India
Overview
This recording recaps an H2O World India hackathon on forecasting air pollution, where more than 250 teams used real-time Central Pollution Control Board data to predict AQI values as a time series problem. Principal data scientist Parul explains why H2O chose the theme, how the competition ran on the in-house H2O Olympics platform built on H2O Wave, and how judging went beyond leaderboard rank to include completeness and storytelling. The session announces the top three teams and reviews their H2O Wave dashboards. It's a quick look at how H2O runs community hackathons and what a strong submission looks like.