The aim of this study is to assess safety conditions in underground stations and surrounding areas where individual trips take place. The study is based on four independent sources: Police crime data, Veolia’s database, SL’s database and Stockholm Safety Survey. Spatial data analysis in GIS underpin the methodology employed combined with detailed fieldwork at the underground stations. Findings show that a relatively small share of reported events are crime, most events are public disorder and disturbance. Clear temporal and spatial variations of both crime and events of public disorder were found. Events tend to happen more often in the evenings-nights, during holidays and weekends and, at least for violence, in the cold months of the year. Despite the fact that underground stations and their surrounding areas are criminogenic places (around 60 per cent of all reported offences to the Police in Stockholm city takes place up to 500 meters from an underground station in 30 per cent of its area), people declare to feel relatively safe there. They are more concerned about safety on the way from/to these transport nodes. T-Centralen might concentrate the highest number of events in Stockholm but it does not keep its top position after events are standardized by daily passengers flow. The so-called “end-stations” often show higher rates of events (crime and public disorder) than stations located in the inner city areas (exceptions are Medborgaplasten, Skanstull and T-centralen for thefts, for instance). Hjulsta, Farsta Strand and Hagsätra show high rates regardless crime type. Some of these stations are also perceived as unsafe. The environment of underground stations follows some common standards (e.g., illumination, gates, real time train arrival time tables, and platform/lounge structures) but they are far from being homogenous. Differences in the environment of underground stations and their contexts have an impact on the stations’ vulnerability to crime and perceived safety. Evidence shows that features that indicate barriers to formal and informal social control are related to higher rates of offences, such as few people in the station, objects hindering visibility/surveillance, corner, hiding places. Good illumination, less presence of physical and social disturbance is often related to lower rates of crime and events. The context of these stations is also important to the stations’ vulnerability. Stations are often more targeted by crime and disorder when they are located in more peripheral neighbourhoods with higher housing instability and population density and fewer police stations. Factors behind crime and disorder at the station are not always relevant to explain perceived safety. Stations perceived as unsafe are often associated with visible social disturbance, with low potential for surveillance, where violence and events of public disorder are common and visible. Although crime is an important component of safety, property crimes, such as theft, do not seem to play a role in explaining perceived safety. Unsafe stations are often located in neighbourhoods with social problems and high housing mobility. On the other hand, safe stations are characterized by an effective formal social control in place, they are often smaller (fewer number of platforms and exits), exhibit high potential for natural surveillance and are not necessarily centrally located. This study finalises making suggestions for safety improvements taking into account different types of crimes, their variation over time and space and the perception of safety by different groups. Finally, it also highlights areas in need of further research, among other things, the need of better knowledge on how to implement safety measures with a whole journey approach, taking into account the specificities of stations environments and needs of different groups of passengers.