Léo is a doctoral student at the University of Bristol and at the Turing. He is interested in increasing the useability and impact of data relating to smallholder farmers in the developing world. Léo began by studying an undergraduate degree in Physics. During his undergraduate, he developed a passion for charity work, seizing opportunities to volunteer in local and international projects.
Following a masters in international development, Léo began working for a small team at the International Livestock Research Institute (ILRI) who developed a Rapid Household Multi-Indicator Survey (RHoMIS). This survey is used by development organisations and research institutes to collect data on smallholder farmers on a project-by-project basis whilst also producing an internationally harmonised dataset.
The RHoMIS tool has increased in popularity, being used to interview approximately 28,000 households in Africa, Latin America and Asia. To better deal with the increased rate of data collection, Léo focused on developing systems for automated data-cleaning and analysis. This reduced the time needed to provide the data to the projects who need it and facilitated the harmonisation of the wider dataset.
Through his PhD Léo would like to target three main areas:
- The availability and usability of RHoMIS data
- Open access to data processing systems
- Representativeness of the data and associated findings
The first tranche of the RHoMIS dataset is soon to be made public and so it is important that this highly multivariate data set is useable. Léo hopes to develop a database that will allow researchers and development organisations to efficiently access the data that they need to conduct their work. The processing systems will also be made public, he hopes his time at the Turing will show him the most effective ways for making this system open source.
Finally projects collecting data with RHoMIS often focus on smaller areas (counties or subcounties) making it difficult to produce generalisable findings. By linking the RHoMIS database to other household surveys, geographic datasets and climate data Léo hopes to use multi-level modelling and small area estimation to conduct research with nationally or regionally representative findings.