Check out GRITS at: grits.eha.io
A natural language processing (NLP) and analysis tool that automatically annotates epidemiologic information in unstructured text sources and diagnosis diseases and extracts locations using an advanced machine learning algorithm.
- Submit unstructured text into app from a variety of sources for analysis
- Run GRITS continuously on social media or news feeds with the GRITS API
- Rapidly extract information about disease events in text like case-counts, symptoms, pathogens, transmission types, disease hosts, dates, and locations
- Ranked diagnosis of potential infectious diseases associated the textual data
- Visualize extracted disease event locations on a map
- View disease timeline of dates associated with event cases in your text
- Store and view your previous GRITS analyses
- Compare the disease events in your text over network maps to assess possible disease spread
Infectious diseases pose a significant threat to global health and economic stability. Due to extensive globalization and urbanization, infectious diseases can spread at unprecedented rates. Early detection of emerging threats is critical to implementing effective responses, and is achievable through robust global disease surveillance. Despite international awareness of the importance of disease surveillance in mitigating infectious disease threats, there are gaps in the current global biosurveillance infrastructure.
The GRITS application is a dynamic web app for performing detailed analysis of text. GRITS analyzes textual data sources like online news outlets, ProMED reports, and blogs. GRITS extracts and succinctly visualizes critical public health information to suggest possible associated infectious diseases faster than traditional biosurveillance methods. GRITS identifies and annotates critical epidemiologic information (e.g., case-counts, symptoms, pathogens, transmission types, hosts, dates, geoname locations) from text and delivers a list of possible diseases associated with submitted text ranked by probability. Then, GRITS visualizes infectious disease occurrence for analysts on a timeline and map, and provides links to similar and potentially related infectious disease reports using a customizable search feature. GRITS is available through a web-application for detailed analysis of specific text, and through an API for rapid analysis of large collections of text, and integration with other applications. The GRITS API can be used to continuously analyze information feeds and large collections of data and can be easily incorporated into larger surveillance systems. GRITS is a flexible and pluripotent tool that contains robust Natural Language Processing (NLP) and machine learning algorithms that can be modified to conduct sophisticated report triaging, expanded to include customized alert systems, or tailored to address other surveillance needs.
The GRITS app is currently in development by the EcoHealth Alliance Research Technology Team and is funded by the Defense Threat Reduction Agency.