Civic Space Data & Forecasts

Data and Forecast Update Reports by Country
  • 7/18/24: India, Indonesia, Kazakhstan, Sri Lanka, Moldova, Serbia, Turkey, Philippines
  • 7/10/24: Albania, Belarus, Georgia, Hungary, North Macedonia, Uzbekistan, Algeria, Liberia, Senegal, Bangladesh, Cambodia
  • 7/4/24: Armenia, Timor Leste, Pakistan, Nicaragua, Malawi, El Salvador, Colombia
  • 6/27/24: Honduras, Ecuador, Kenya, Kyrgyzstan, South Africa
  • 6/20/24: Benin, Burkina Faso, Cameroon, Zambia, Zimbabwe, Azerbaijan
  • 6/1/24: DR Congo, Ethiopia, Guatemala, Indonesia, Jamaica, Mali, Mauritania, Niger, Nigeria, Tanzania, Rwanda
  • 5/10/24: Belarus, Hungary, India, Malaysia, Moldova, Sri Lanka, Uzbekistan, Angola, Cameroon, Ghana, Indonesia, Morocco, Mozambique, Nepal, Paraguay, Peru, South Sudan, Tunisia, Uganda
  • 4/24/24: Albania, Armenia, Bangladesh, Cambodia, Georgia, Kazakhstan, Kosovo, Macedonia, Namibia, Philippines, Serbia, Turkey, Ukraine
  • 4/1/24: Azerbaijan, El Salvador, Honduras, Kyrgyzstan, Liberia, Malawi, South Africa
  • 3/20/24: Algeria, Benin, Burkina Faso, Colombia, DR Congo, Ecuador, Ethiopia, Guatemala, Jamaica, Kenya, Mali, Nicaragua, Senegal, Tanzania, Zimbabwe
  • 2/28/24: Angola, Cameroon, Ethiopia, Ghana, Mauritania, Morocco, Mozambique, Niger, Nigeria, Pakistan, Paraguay, Peru, Rwanda, South Sudan, Tunisia, Uganda
  • 1/31/24: Armenia, Hungary, India, Indonesia, Moldova, Namibia, Nepal, Serbia, Sri Lanka, Uzbekistan, Zambia
  • 1/12/24: Albania, Belarus, Georgia, Kazakhstan, Kosovo, North Macedonia, Turkey, Ukraine, Bangladesh, Cambodia, Philippines, Malaysia
  • 12/18/23: Colombia, El Salvador, Guatemala, Honduras, Algeria, Liberia, Malawi, South Africa, Kyrgyzstan
  • 11/25/23: Pakistan, Angola, Kenya, Mali, Nigeria, Senegal, South Sudan, Azerbaijan, Nicaragua, Peru
  • 11/14/23: DR Congo, Ethiopia, Ghana, Mozambique, Rwanda, Mauritania, Niger, Tunisia, Ecuador, Paraguay
  • 11/5/23: Benin, Cameroon, Zimbabwe, Tanzania, Uganda, Armenia, Jamaica, Morocco
  • 10/22/23: Indonesia, Malaysia, Sri Lanka, Uzbekistan, Kosovo
  • 10/8/23: Bangladesh, Cambodia, Georgia, Kazakhstan, Macedonia, Serbia
  • 9/22/23: Albania, Hungary, Moldova, Turkey, Ukraine, India, Philippines
  • 9/10/23: Liberia, Colombia, Honduras, Guatemala, Algeria, Malawi, Zambia, South Africa
  • 8/30/23: Mali, Senegal, Kenya, Kyrgyzstan, Nigeria, El Salvador, Nicaragua, Peru
  • 8/9/23: South Sudan, Niger, Ethiopia, Cameroon, Mozambique, Rwanda, Morocco, Tunisia, Kazakhstan
  • 7/27/23: Algeria, Armenia, Paraguay, Benin, DR Congo, Ghana, Tanzania, Uganda, Zimbabwe, Angola, Mauritania
  • 7/13/23: Georgia, Hungary, Kosovo, North Macedonia, Indonesia, Bangladesh, Uzbekistan
  • 6/28/23: Albania, Cambodia, Malaysia, Philippines, Serbia, South Africa, Sri Lanka, Turkey, Ukraine
  • 6/6/23: El Salvador, Guatemala, Jamaica, Nicaragua, Kyrgyzstan, Mali, Zambia
  • 5/31/23: Kazakhstan, Moldova, Peru, Cameroon, Kenya, Malawi, Rwanda, Senegal, Zimbabwe, Colombia, Ecuador, Honduras, Paraguay, Uzbekistan
  • 5/16/23: Angola, Benin, Ethiopia , Ghana, Mauritania, Tanzania, India
  • 4/28/23: Azerbaijan, Armenia, Indonesia, Malaysia, Niger, Morocco, Nigeria, DR Congo, Uganda, Tunisia, Mozambique, El Salvador
  • 4/24/23: Kyrgyzstan, Cambodia, Bangladesh, Hungary, Philippines, Sri Lanka, Turkey
  • 4/10/23: Georgia, Ukraine
  • 3/22/23: Albania, Kosovo, Serbia, Tunisia, Nicaragua
  • 3/4/23: Colombia, Ecuador, Guatemala, Honduras, Paraguay, Kenya
  • 2/24/23: India, Mozambique, Malawi, Jamaica, South Africa, Turkey, Zambia
  • 1/31/23: Serbia, Kosovo, Nigeria, Morocco, Niger, Mauritania, Malaysia, DRC, Cameroon, Albania, Cambodia
  • 1/11/23: Paraguay, Benin, Zimbabwe, Tanzania, Sri Lanka
  • 12/17/22: Uganda, Guatemala, Angola, Ghana, Philippines, Uzbekistan
Recent Pipeline Updates
  • February 28, 2024: Since the last update for these countries, we have implemented a new shock detection algorithm to identify major events. This new algorithm integrates statistical and machine learning methods to identify major jumps in reporting on each event. This new approach is better able to detect events happening on the ground from disturbances in the volume of reporting. We have also integrated a new method of summarizing major events detected in our data since the last update. This new method uses GPT4 to summarize all articles from our database reporting on an event in months when we detect a shock. However, we use human supervision to ensure these AI summaries are accurate.
  • March 21, 2023: We have recently applied a new classification system to a subset of our event variables. This classification removes articles that report on events that are not directly relevant to civic space. For example, articles reporting on arrests or legal actions that involve regular criminals rather than important political figures are no longer included in these event categories. This means that our measures for the affected categories will be significantly lower than they were in previous months. For this reason, the forecasts made in past months are not directly comparable to the actual data. The affected categories are: Arrests, Cooperation, Corruption, Defamation Cases, Legal Actions, Legal Changes, Purges, Raids, Political threats, Lethal violence, and Non-lethal violence.
  • December 17, 2022: We have recently completed a review of all local sources that we are scraping to assess their alignment/bias. When we have a larger volume of news from state-aligned sources, we apply weighting that over-represents news from more independent sources. Currently, we apply weighting to DR Congo, Malaysia, Zambia, Sri Lanka, India, Angola, and the Philippines.

Civic Space Early Warning System (CSEWS) Launch Event

On February 9, 2023, the Machine Learning for Peace (MLP) project held a launch event for their new Civic Space Early Warning System (CSEWS). Following introductory remarks by Rosarie Tucci, the Director of USAID’s Center for Democracy, Human Rights, and Governance, the MLP team introduces the CSEWS and provides an overview of recent research findings relevant to USAID’s work on democracy and governance around the world.

Evidence for Impact Webinar Series

On May 25, 2023, the Machine Learning for Peace (MLP) was featured in the Promise and Perils of Data Science for Development webinar as part of the Social Impact Evidence for Impact Webinar Series. The webinar discussed how data science can be used to improve development effectiveness. The presentation concluded with critical reflections from within USAID on the promise and pitfalls of using data science for development.

Short Video Guide to MLP Data and Forecast Dashhboards

For an introduction to our data, the forecasts, and the dashboards linked above, please view this short Tech Demo to featured at the Global Digital Development Forum in May 2022. This video provides the basic information necessary to navigate through the website and interact with the data and forecasts.