Data API
Programmatic access to PaCE forecast data. Updated automatically on the 1st of each month.
Country Forecasts
Quick Start
Fetch the latest forecasts in your preferred language. All endpoints are stable and updated monthly on the 1st.
import pandas as pd
import requests
# Fetch latest forecasts and historical data
forecasts = pd.read_csv("https://forecastlab.org/data/forecasts/latest/forecasts_h12.csv")
historical = pd.read_csv("https://forecastlab.org/data/forecasts/latest/Hist.csv")
metadata = requests.get("https://forecastlab.org/data/forecasts/latest/metadata.json").json()
print(f"Forecast period: {metadata['forecast_start_date']}")
print(forecasts["Ukraine"].head()) # First 5 months for UkraineFor full API specifications, see the Complete API Reference.
Latest Forecasts
Historical Data
Download the historical training data used to generate the predictions:
Historical Archive
Available from 1989-01 to 2026-04.
Prediction Bounds: Add _min or _max to forecast filenames: forecasts_h12_min.csv, forecasts_h12_max.csv. What they represent: Empirical bounds derived from historical analogues. The model identifies past conflict trajectories similar to the present situation, then uses the range of those outcomes to estimate lower and upper prediction bounds. Not statistical confidence intervals, but scenario-based uncertainty estimates.
Archive Downloads
Access any historical forecast period from 1989-01 to 2026-04.
| Period | Bundle | Hist.csv | metadata.json | h6 files | h12 files |
|---|---|---|---|---|---|
| 2026-04 | forecasts-2026-04.zip | Hist.csv | metadata.json | ||
| 2026-03 | forecasts-2026-03.zip | Hist.csv | metadata.json | ||
| 2026-02 | forecasts-2026-02.zip | Hist.csv | metadata.json | ||
| 2026-01 | forecasts-2026-01.zip | Hist.csv | metadata.json | ||
| 2025-12 | forecasts-2025-12.zip | Hist.csv | metadata.json | ||
| 2025-11 | forecasts-2025-11.zip | Hist.csv | metadata.json |
Sub-national Forecasts (PRIO-GRID)
High-resolution spatial forecasts on a 0.5° grid (~55 km cells). Download forecasts for period 2026-03 in CSV or GeoJSON format.
Quick Start
import pandas as pd
import requests
# Download grid forecasts
grid = pd.read_csv("https://forecastlab.org/data/grid/2026-03.csv")
geojson = requests.get("https://forecastlab.org/data/grid/2026-03.geo.json").json()
print(f"Grid cells: {len(grid)}")
print(grid.head())CSV Files
GeoJSON Files
License and Citation
License
CC BY-NC 4.0. For commercial use, contact us.
Update Schedule
Forecasts generated on the 1st at 01:00 UTC. Data available by 03:00 UTC.
Citation
Schincariol, T., Frank, H., & Chadefaux, T. (2025). JPR. DOI