Forecasts
INSPIRES is testing how machine learning can help predict changes in civic space before they occur in select countries around the world. DevLab has created an enormous database and tools to online news for events relevant to civic space and attempts by authoritarian regimes to influence civic space abroad, referred to as Resurgent Authoritarian Influence (RAI). Each month, we collect thousands of news articles posted online from international, regional, and domestic news sources. We use machine learning to read and classify each article according to 19 civic space event categories and 22 RAI event categories. Data scientists combine data on these events with high frequency economic data to forecast changes in civic space, including a ‘Civic Space Index’ that summarizes movement across many civic space events. You can learn more about this process in our Pipeline Report.
This web application allows users to view forecasts about levels of civic space activity in the coming months for each country where we have data. Using the application below, you can select a country and event of interest, generate a forecast using our data and forecasting models, and see a basic description of the results. While these predictions will not always come true, INSPIRES regularly assess their accuracy to improve our forecasts. We regularly produce more detailed forecasting reports for specific countries. If you would like to access a more detailed forecasting report for a specific country, please see the contact information provided at the bottom of this page.