Flu nowcasting system.
Note: This system is designed to nowcast ILI driven by seasonal influenza and is NOT designed to nowcast ILI during the COVID-19 pandemic.
Delphi’s “Wisdom of crowds” forecasting system: Used for Chikungunya, flu and most recently Covid.
Note: This system is has been repurposed to forecast ILI during the COVID-19 pandemic.
State-level weekly forecasts of ILI (influenza-like illness)
Note: This system is designed to forecast ILI driven by seasonal influenza and is NOT designed to forecast ILI during the COVID-19 pandemic. We have temporarily shut it down to focus on COVID-19.
We have participated, and have done very well, in all epidemiological forecasting challenges organized by the US government to date:
by CDC 2013 – current
Epidemiological time series visualizer
API for getting up-to-date epidemiological data (also available via a web interface through EpiVis)
Epidemiological forecasting R package
Epidemiological nowcasting via sensor fusion
Visual comparison of scored submissions to CDC Flu Forecasting Challenge (provide your own score files)
Epidemiological modeling utilities (e.g., date/epi-weeks conversions)
All source code is freely available on GitHub.
Delphi’s notable achievements
We’ve welcomed 13 Google Fellows to Delphi and are excited to have them on board.
We launched our COVIDcast system, which displays indicators related to COVID-19 activity level across the U.S. These indicators are derived from a variety of anonymized, aggregated data sources made available by multiple partners, and are publicly available at the COVIDcast endpoint of our Epidata API.
Related news articles:
We are focusing our efforts at this point on COVID-19 nowcasting and forecasting. We are adapting our existing systems, and developing new ones. Some of our regular activities may be halted as a result.
CDC has just named us “National Center of Excellence for Influenza Forecasting” (one of two nationally.)
and yet again! Our forecasting systems took the top spot each of the three separate flu forecasting challenges of 2017-2018 (out of up to 30 submissions). (Results summary.)