23 Jul 2020, 13:25

Zindi solutions will help Malawi prepare better for flooding

To help Malawi better predict and prepare for major flooding events, Zindi teamed up with Arm and UNICEF in a double challenge to use machine learning and data visualisation to help solve the problem.

On 14 March 2019, Tropical Cyclone Idai made landfall at the port of Beira, Mozambique, before moving west to become Africa's deadliest tropical cyclone ever. In the aftermath, millions of people in Malawi, Mozambique and Zimbabwe have lost homes, livelihoods and food security.

To help prevent this happening again, Zindi hosted the UNICEF Arm 2030 Vision #1: Flood Prediction in Malawi and UNICEF Arm 2030 Vision #2: Malawi Floods Data Visualization and Reporting challenges, which attracted more than 1600 data scientists to compete for a total prize money of USD15 000.

The flood prediction challenge asked data scientists to build a machine learning model to help predict the location and extent of floods in southern Malawi, using data from flooding in 2015.

The winners of this challenge were:

The data visualisation challenge asked Zindians to interpret and generate insights from the 2015 flooding data and its aftermath in southern Malawi, by creating an original data visualisation or written report that explores the impact of the 2015 floods. A judging panel reviewed the usefulness and impact of the reports, as well as how the findings were communicated and how novel those findings were.

The winners of this challenge were:

The three winning reports can be seen here.

Both competitions were sponsored by Arm and UNICEF as part of 2030Vision, an initiative to drive technological solutions to the Sustainable Development Goals (SDGs). UNICEF is currently assessing the models and reports before further action is taken to put these solutions into practice.

Thank you to everyone who made a submission in these competitions, and congratulations to all our winners!