Data Analytics & Policy Research

Data becomes evidence only once it has been rigorously analysed and translated into something decision-makers can act on. We work with complex, often messy, real-world development data — administrative records, survey datasets, panel data, and programme monitoring data — and apply advanced statistical and econometric methods to generate insights that are both technically robust and genuinely useful to policy.

Our analytical toolkit spans econometric modelling for causal and correlational analysis, predictive analytics for forecasting and targeting applications, and mixed-methods approaches that combine statistical analysis with qualitative and contextual understanding. We work across cross-sectional, longitudinal, and panel data structures, and we bring particular attention to the data quality issues — missingness, measurement error, attrition — that are the norm rather than the exception in development data, and that can quietly undermine analysis if left unaddressed.

Our policy research practice sits at the intersection of this analytical capacity and a deep understanding of the policy process. We support governments and development partners not only in generating findings but in working through what those findings mean for programme design, resource allocation, and scale-up decisions. This includes cost-effectiveness and economic evaluation work, which helps partners weigh the value of a given intervention against its costs and against alternative uses of the same resources — often a decisive factor in scale-up decisions even when an intervention has demonstrated positive impact.

We aim to make our analysis accessible without oversimplifying it. Policy research that is technically sound but inaccessible to non-specialist decision-makers has limited influence; policy research that is accessible but analytically weak risks misleading them. We invest deliberately in the space between these failure modes — clear technical reporting, well-designed data visualisation, and direct engagement with decision-makers to walk through what findings do and do not support.

This work often draws directly on datasets generated through our own survey research and evaluation work, but we also engage with existing administrative and secondary datasets where partners need analytical capacity applied to data they already hold.