We have considerable experience designing algorithms and models on behalf of clients and funders. Some of our projects included building poverty forecasts for difficult environments, creating disease prevalence and birth models at both the regional and global scale, and middle class growth estimates for India and Latin America.
World Data Lab's models are built using a variety of methods, including: Bayesian Model Averaging (BMA), Convolutional Neural Network (CNN) training, Categorical Luminosity Downscaling, Twinning, Random Forest, lifetable estimation, and Beta-Lorenz curve forecasting. Our use of these and other modeling and analytical frameworks is overseen and supported by a team of advisors from both academia and Silicon Valley.
Our custom models benefit from rich data sources. As a strategic partner of the OECD's PARIS21 organization, World Data Lab enjoys strong collaborative relationships with many National Statistical Offices worldwide. Additional strategic partnerships with the International Institute for Applied Systems Analysis (IIASA) and GeoVille ensure that our models are supported by weather observation data (including satellite imagery), and other more experimental data sources.
TRANSPARENCY, NOT BLACK BOXES
Our global team of data scientists, eonomists, and demographers includes some of the world's leaders in their respective fields. We are proud of our models and methods. The support we have received from leading public and private organizations speaks for itself. Partnering with World Data Lab to develop a new model means never having to worry about the integrity or credibility of data.
Expertise to Work
With Your Data
We offer a full range of services from model refinement and integration of additional and proprietary datasets to complete design, data sourcing, coding, and delivery.