Geospark Analytics tuned the Hyperion Health Event machine learning model to identify, track and analyze events associated with the spread of COVID-19. The data includes information from news and social media. We released the results of our Health Event model for free. The data tracks global reporting about the virus and have been accessed more that 15,000 times in the first month and has been integrated into other applications providing a unique view into the world’s COVID-19 information.

Geospark Analytics' Events are derived from an AI-driven deep learning model that categorizes disparate forms of information into classes of activities. While most systems categorize events based on a keyword-driven taxonomy or ontology, Hyperion Events learn patterns in unstructured data to recognize patterns in social media, news media, and other source data and categorize it automatically into event themes. This feature services is derived from that deep learning model tuned to identify the most relevant news related to COVID-19.

Experience Geospark Analytics AI data in our interactive Story Map.

Data is available as a hosted Feature Service to easily integrate with all ESRI products in the ArcGIS ecosystem.

For more information on Geospark Analytics Event models visit our product page.

For use cases on how Geospark Analytics Event models can be used visit our blog.

Contact us at sales@geospark.io for multiuser or enterprise pricing.

Event Theme	Periodicity	Sources
Disaster Data	Near-real-time	News, Social, Government Sites
Crime Data	Near-real-time	News, Social, Bespoke Crime Data Sources
Conflict Data	Near-real-time and daily	News and Social
Travel Safety Data	Near-real-time, daily, long-term	Governments (US, UK, AUS, NZ)
Unrest Data	Near-real-time, daily, long-term	News, Social, Government Sites
Health Data	Near-real-time	News, Social
Terrorism Data	Near-real-time, daily	News, Social
Transportation Data	Near-real-time, daily	News, Social