Gro launches new Real-Time Assessment on China’s Food Security

Learn More

Summary & Output

This tool offers users select frameworks for discovering anomalous weather events that could have significant impact on crop production. The example code analyzes data series for a limited number of important agricultural regions, but could be easily edited to detect even more types of anomalies for any series in any region.
This code could be used to automatically check a large amount of data and alert users whenever a significant anomalous event occurs.

Attributes

  • Discover meaningful signals by finding anomalous values in large amounts of data.
  • Edit the data, regions, or significance testing to fit series of greatest interest to you.
Accessibility: The code for Gro’s framework is functional exclusively with the Gro API Client.

Inputs

This tool can be edited to work with almost any time series data. However, it has proven to be particularly useful for analyzing weather and other satellite-derived data. The provided example code works with the following sample sources in a limited number of regions, but the code can be edited to work with other sources in any region:
LST

Daily land surface temperature for districts in Argentina’s Salta province

NDVI

8-day averages of anomalies for districts in Brazil’s Tocantins state

Methodology

Gro: How does anomaly detection tool works
  • Retrieve all time series data for selected regions
  • In the case of daily LST, transform data into 7-day averages
  • Check if the most recent data points are greater than two standard deviations from the mean
  • Format into a text alert

Notebook

Contact sales