2013 – 2016 · Institute for Health Metrics and Evaluation
Measuring the health of 188 countries
I started as a data analyst on the Global Burden of Disease study, where estimates for every country had to hold up to peer review. One of the papers I contributed to tracked deaths of children under five from 1990 onward.
Each dot is a point of land, raised and colored by its country's under-5 mortality. Keep scrolling and the years run forward: worldwide, the rate has fallen by more than half since 1990.
Data: UN IGME via Our World in Data, 1990 to 2024. IHME's own estimates can't be redistributed, so this uses the open UN series.
2017 – 2021 · Kaiser Permanente
Seeing a surge before it arrives
At Kaiser I worked on analytics for 20 hospitals serving 4.5 million members. When COVID-19 arrived I designed the central forecasting dashboard, using anomaly detection on call-center data to predict surges four to six weeks ahead.
Keep scrolling and the dots trace weekly COVID-19 cases in the United States. A red line appears ahead of each surge, where an early warning would fall: the head start that lets a hospital staff up before the wave arrives.
Data: WHO via Our World in Data. Each red line sits five weeks before a wave reaches half its height. They illustrate the idea and are not Kaiser's forecasts.
2021 – 2022 · Amazon Care
Care for people wherever they live
As product and client analytics lead for Amazon's telehealth service, I modeled health-record, operational and financial data to guide decisions on acquisition, engagement and retention. That work contributed to 300,000 new enrollments and a 120% rise in engagement.
Telehealth reaches people where they are, so the dots now show where Americans live. Each one stands for about 42,000 residents, and together they look like the country's lights seen from orbit.
Data: 2020 Census population centers for every census tract. The map shows the US population, not where Amazon Care members were.
2023 – now · Uber
Organizations as data
Today I build analytics products for people and corporate operations: tools that show leaders the shape of their organization and flag when its metrics shift.
Uber's data is confidential, so these dots are the employees of Kestrel, a synthetic company I simulated. Green dots are managers and orange dots are executives.
Open Org Health Explorer →Data: simulated. The same company powers Org Health Explorer.