Welcome to the Wikimpacts project!
A database of climate impacts inferred from Wikipedia mining with large language models
Data quality disclaimer
The Wikimpacts project is an ongoing initiative, and data is constantly being updated. The data is provided "as is" without guarantees of accuracy or completeness. Users should be aware that it may contain errors or omissions and interpret it cautiously. The project team assumes no responsibility for any risks or losses from the use of this data.
More than 3,173 Wikipedia articles processed, covering 2,726 events from 1034 to 2024, , with 17,912 national and 32,343 sub-national entries
The Wikimpacts project aims to address the challenges posed by extreme weather events, such as floods, heatwaves, wildfires, and hurricanes, which increasingly impact societies and cause significant economic damage due to climate change. Current databases are limited by incomplete data and sampling biases. Wikimpacts overcomes these limitations as a new database developed with natural language processing and machine learning to mine data from Wikipedia and other online sources, capturing detailed information on extreme weather events, including their timing, location, and socioeconomic impacts.The Wikimpacts database will enhance our understanding of the links between climate change and extreme weather, potentially providing unprecedented insights into the role of anthropogenic climate change in driving economic damages and loss of life globally.
We introduce Wikimpacts 1.0, a new global database on climate impacts developed using natural language processing techniques. Impact data for each event is recorded at three levels: event, national, and sub-national. Categories include the number of deaths, injuries, homelessness, displacements, affected individuals, damaged buildings, and insured or total economic damages.
Extreme climate events like storms, heatwaves, wildfires, floods, and droughts pose serious threats to human society and ecosystems. Impact database can be used in impact forecasting, early warning, and disaster risk management.
This database utilizes the GPT4o large language model for extracting information, following document selection, post-processing, and data consolidation. In this release, we have processed 3,173 Wikipedia articles. This dataset encompasses 2,726 events from 1034 to 2024, featuring 17,912 national and 32,343 sub-national data entries.
Wikimpacts in numbers
About 50% of climate impacts reported in the past XX years
Total Homelessness
Total Damage
Total Insured Damage
Total Building Damage
Total Deaths
Total Injuries
Total Displacement
Multi-scale focus
The Wikimpacts database accounts for climate impact records ranging from local and national to regional and global scales.
- Most impacted countries
- Number of fatalities
- Predominant climate impacts
- Trends and many more...
