How to Use & Methodology
For a quick user guide, click here.
What is included in the monitor?
The monitor provides an automatically generated overview of media articles on incidents of firearm violence and firearm seizures published in 369 online media sources in all 27 EU Member States. Importantly: police shootings and suicides with a firearm are not included in the definition of what constitutes firearms violence.
These sources are reported on the level of the incident: articles on the same incident are therefore clustered together. The articles are geographically located on a map, with each dot representing one incident. The search bar allows to limit the search to specific time periods, countries and contexts. The articles and the contextual information extracted from them, can be accessed via the feed button. The contextual information, generated by the AI-algorithm, is displayed in the blue box next to the article.
Limitations related to the use of media articles as source
The use of media articles as a source of information has some important limitations all users should be aware of:
- Media articles do not provide a comprehensive overview of firearms violence incidents in a country. They can therefore not be used as an objective indication of the prevalence of firearms violence in the selected country/ countries.
- The information in media articles is not fact checked and as such may not always be 100% correct. We do not take responsibility for any mistakes or incorrect wrong information included reported in the articles. Measures (e.g. regular manual follow-up or triangulation of different media sources) are taken to increase the reliability and validity of the information included in the monitor, but complete accuracy cannot be guaranteed.
How are the articles identified?
The articles are automatically identified, selected, clustered and analyzed using different Artificial Intelligence techniques. Despite substantial efforts invested in optimizing the reliability of these AI instruments, mistakes, such as non-relevant articles could be included, articles on the same incident are not clustered together resulting in double countings, or incorrect contextual information is included in the context box, could still occur.
The absolute numbers of incidents reported in and visualized on the map should therefore not be taken at face value. Comparisons over time and between countries should be done with great caution.
If you see any mistakes (e.g. wrong articles, articles wrongly or not clustered together, incorrect contextual information), please let us know via our contact page.
Want to know more about the methodology behind the monitor? Consult the methodological report here.
A CSV-file including a selected set of articles and the different categories included in the contextual box could be made available for research or strategic reasons upon request. If interested, please send a motivated request using our contact page.
Methodological Report