Portrait of Mattéo Larrodé

Mattéo Larrodé

Data Analyst @ the Zambia Evidence Lab
MSc Social Data Science @ the University of Oxford

I aim to use data science for social good. I'm especially drawn to how machine learning and AI can strengthen humanitarian decision-making, and improve the lives of vulnerable people.

So far, I have built data tools for the British Red Cross and Zambian government agencies, and used satellite imagery to track building destruction in Gaza. I want to take this further, towards the questions humanitarian organisations face: where displaced people are and what they need, how to develop robust early warning systems and limit the impact of disasters when they hit, and how to document human rights abuses.

OSINT, investigative data journalism and innovative data storytelling are angles through which I approach these questions. I draw a lot of inspiration from organisations at the cutting edge of these fields: Forensic Architecture, Lighthouse Reports and the Human Rights Data Analysis Group are some of them.

Research & Open-source

As a strong open-source enthusiast, I make my research, and personal projects as transparent and reproducible as possible. Here is a little selection!

Monitoring of Building Destruction in Gaza

Detecting building damage in Gaza using public Sentinel-1 data with machine learning.

Wildfire Risk and Social Vulnerability

The wildfires R package couples wildfire risk (modelled via random forest) with social vulnerability in the UK. Paper under review in the International Journal of Disaster Risk Reduction.

Health Inequalities Explorer

Open access web app to explore location based health inequalities across different UK geographies.

DEPAHRI

An index of how hard it is to physically and digitally access healthcare in England, Wales and Scotland. Source code and data in the DEPAHRI R package.

"Balance Ton Rap"

Quantifying gender stereotypes in 8,000 French rap songs using word embedding association tests.

A Peaky Blinders Network Analysis

An introductory network analysis of the Peaky Blinders TV show, using web-scraped transcripts of all 36 episodes.

Experience

See CV

Education

Conferences & Webinars

Recent events I have attended to keep learning about OSINT and investigative data work.

Get in touch