Framing Gaza

The articles were collected using a combination of the open-source package News Please and the Media Cloud project, which allowed for safe and systematic gathering. The sample was drawn from the United States, the United Kingdom, France, and Germany, based on clear criteria: the platforms had to be English-language, widely used according to SimilarWeb, and they had to publish a significant number of articles on Gaza. The dataset includes only materials available online, noting that some articles could not be collected due to technical issues or lack of availability on publishers’ websites.

The analysis relied on Association Rule Mining (ARM), implemented through the Apriori algorithm. Titles were analyzed separately, with lower thresholds (support 1%, confidence 10%) to capture rare but meaningful framing patterns in shorter text. Full articles were analyzed with stricter thresholds (support 15%, confidence 70%) to reduce noise and ensure stronger recurring patterns. In both cases, only rules with a lift value greater than 1 were retained, to guarantee that the relationships between words were significant rather than random.

Yes, the dataset is accessible to all journalists, researchers, and activists interested in the topic, however, we prefer that personnel who is interested in accessing the dataset, to take a workshop with us, to learn what they can about ARM analysis and Apriori, before start mining in our datasets. It’s not mandatory, but we strongly encourage it. If you are interested, please reach out via email: info@anmat.media
or using this form:
https://anmat-media.github.io/anmat-workshop/

The datasets is available on Kaggle, you will be able to access it from here, or from Anmat account on Kaggle