We do data science, program evaluation, and applied AI for public and mission-driven organizations.
Selected work
Predictive models, classification, clustering, and deep learning, plus web scraping and pulling structure out of messy data. From a first exploratory look to pipelines you can run in production.
Program evaluation and causal inference for public and social policy: A/B tests, randomized trials, and quasi-experimental designs.
Building and stress-testing LLM and agent-based tools, and auditing automated decision systems for bias and real-world impact.
Interactive dashboards, publication-quality graphics, and full data applications, front end to database.
Spatial statistics, remote sensing, and GIS: satellite imagery, environmental monitoring, and infrastructure mapping.
Quantitative analysis with Gregg Colburn demonstrating that housing-market conditions (not individual factors) explain regional variation in U.S. homelessness rates. Informed policy debates at the municipal, state, and federal levels.
Teaching statistical learning and machine learning to the human-services data community through hands-on practitioner workshops, including at the National Human Services Data Consortium. The applied thread uses these methods to surface racial disparities in how coordinated-entry systems prioritize people experiencing homelessness.
Interdisciplinary alliance of neurologists, cognitive scientists, clinicians, and epidemiologists advancing research at the intersection of climate change and brain health. Affiliated with Columbia Mailman School of Public Health.
A field guide to building and funding a data team in a small or nonprofit newsroom, covering hiring, tools, responsible AI, and fundraising. A Caldern project with the Reynolds Journalism Institute at the Missouri School of Journalism.
We work with public agencies, nonprofits, foundations, and mission-driven companies. We're based in Seattle and work remotely with teams anywhere.
The firm is led by Clayton Aldern, a data scientist and policy researcher.