The growth of large, open datasets coupled with an acceleration of technical developments, including GIS solutions, opens the door to new challenges in transport research. One of the emerging fields of research is the temporal dynamics of accessibility. The increase in availability of General Transit Feed Specification (GTFS) data permits the inclusion of very detailed, schedule-based travel time information. In the study presented we focus on the spatial and temporal variation in accessibility by public transport in the city of Szczecin (Poland). This paper advocates the necessity of incorporating a temporal component in accessi- bility analysis. We conducted a full day analysis for 1 day using averaged 15-min-long time periods at a very detailed spatial scale (enumeration districts). Based on the calculated origin-destination matrix in 96 time-profiles we calculated the potential accessibility indicator. Then we investigated spatial disparities and their variability during the day-long observation. Apart from the well-known spatial disparities in accessibility level, our findings underline the uncertainty of the accessibility pattern. Moreover, the results show that less accessible areas are also more affected by the daily variation in accessibility level. The findings provide a more realistic insight into accessibility patterns which will be useful for transport planners and policy makers.