19 Jul 2019, 10:33

Meet the winners of the Mobile Money and Financial Inclusion in Tanzania challenge

Get insights from challenge winners! A special thank you to the winners for their generous feedback.

Zindi is excited to announce the winners of the recently closed challenge that focussed on financial inclusion in Tanzania. The objective of the competition was to create a machine learning model to predict which individuals were most likely to use mobile money and other financial services (savings, credit, and insurance). The challenge attracted 447 data scientists from across the continent and around the world, of whom over 163 were on the leaderboard. We are happy to introduce the winners of the competition: Nikita Churkin of the Russian Federation and Raheem Nasirudeen Adeleye of Nigeria!

Name: Nikita Churkin (1st prize)

Zindi handle: NikitaChurkin

Where are you from? Russian Federation

Tell us a bit about yourself.

I have bachelor's degree in Economics and two Master's degrees: one in Mathematics and the other in Computer Science. I currently work in the financial sector.

Tell us about the approach you took.

I used strong validation scheme and target decomposition (solving two binary tasks instead of one multiclass task).

What were the things that made the difference for you that you think others can learn from?

The key was not to overfit to a small dataset.

Nikita's score on the private leader board was 0.7074 however his score on the public was 0.7116 and he was in 20th place. This is a testament to him not overfitting to the public leaderboard and thinking about the bigger problem.

What are the biggest areas of opportunity you see for AI in Africa over the next few years, and what are you looking forward to for the Zindi community?

There are many unobserved datasets in Africa that need to be explored. I look forward to learning new tricks to solving data challenges.

Name: Raheem Nasirudeen Adeleye (3rd prize)

Zindi handle: Nasere

Where are you from? Osun State, Nigeria

Tell us a bit about yourself.

I have a National Diploma (OND) from The Polytechnic Ibadan in Mathematics and Statistics. I am also a facilitator at AI Saturdays Ibadan.

Tell us about the approach you took.

I started using the raw data, however this did not give the best model. I decided to enrich the data using ArcGIS. I generated possible features from longitude and latitude. My final model averaged 3 catboost models, each with different parameters.

What were the things that made the difference for you that you think others can learn from?

Understanding the problem. I became familiar with the domain knowledge, used the No Free Lunch model and hard work and dedication.

Raheem took 3rd place with a score of 0.7174 on the public leaderboard.

What are the biggest areas of opportunity you see for AI in Africa over the next few years, and what are you looking forward to for the Zindi community?

I would love to see AI improving health, financial inclusion and sustainable development growth in Africa. I look forward to Zindians solving both local and international problems using machine learning and AI.

What are your thoughts on our winners' feedback? Engage via the Discussions page or leave a comment on social media.

This competition was sponsored by HDIF, Tanzania Data Lab (dLab) and Esri.