This dataset provides a compilation of the data used to analyze and identify potentially dangerousplaces for women. Multiple data collection techniques, including official data downloads, webscraping, and participatory mapping, were combined for integration, applying specific processing.The datasets refer to four cities: Valencia (Spain), Dublin (Ireland), San Francisco (United States),and Toluca (Mexico).Depending on the availability and context of each city, the datasets are classified into threecategories: DATA, TWT, and MAP. The DATA prefix refers to files containing the results of theanalysis of socioeconomic variables downloaded from official sources; for the mapping, thestandard territorial unit was a 25x25 m grid for Valencia and 50x50 m for Dublin and San Francisco.The files with the prefix TWT are composed of datasets containing tweets collected through webscraping and analyzed using natural language processing (NLP) algorithms and neural networks;the purpose is to identify and classify tweets related to gender violence, feelings of fear, orperceptions of insecurity. For MAP files, participants gathered them through participatorymapping processes, using specific calls to public space users and a supporting web applicationdesigned for this purpose.