Fighting Fire with Data Hackathon
Predicting Burned Area in Zimbabwe
60 data scientists enrolled, 23 on the leaderboard
27 June 07:45 (9 hours)

Each year, thousands of fires blaze across the African continent. Some are natural occurrences, part of a ‘fire cycle’ that can actually benefit some dryland ecosystems. Many are started intentionally, used to clear land or to prepare fields for planting. And some are wildfires, which can rage over large areas and cause huge amounts of damage. Whatever the cause, fires pour vast amounts of CO2 into the atmosphere, along with smoke that degrades air quality for those living downwind.

Figuring out the dynamics that influence where and when these fires occur can help us to better understand their effects. And predicting how these dynamics will play out in the future, under different climatic conditions, could prove extremely useful. For this challenge, the goal is to do exactly that. We’ve aggregated data on burned areas across ZImbabwe for each month since 2001. You’ll be given the burn area data up to the end of 2013, along with some additional information (such as rainfall, temperature, land cover etc) that extends into the test period. The challenge is to build a model capable of predicting the burned area in different locations over the 2014 to 2016 test period based on only this information.

Be sure to join us for the webinar associated with the competition! Contact for zoom details or tune into the livestream at The event starts at 9am CAT and will feature some background on the competition and the data that went into it, as well as a tutorial session that will run from 10am CAT onwards after the competition opens.