Mapping public transport time-geographies in Tallinn

An intern at PUTSPACE project, Agata Mielczarek, analysed public transport and social inequalities in Tallinn. 


This video depicts public transport connectivity of the city of Tallinn: how long it takes to get to the city centre from different parts of the city? The city centre is represented in this video by Viru Centre. Of course, it is an arbitrary spot but as it is a major shopping centre and the central bus terminal located between the Old Town and the new CBD of Tallinn, it is a crucial destination for many residents. All the data analyses were made in ArcGIS Pro 2.4.0, using different programme extensions. The tool “Network Analysis – Service Area” was utilised for accessibility analysis, calculating times it takes to get from various points in the city to the Viru Centre. Each of the 144 maps in this video presents accessibility to Viru Centre by public transportation (including also walking distance to the bus stop) and on foot, for every 10 minutes in 24 hours, assuming that the walk speed of pedestrians is 5 km/h. Combining these maps into a short video shows geographical as well as temporal variability that occurs throughout the day. Of course, there are major differences in travel times between late hours or early mornings compared those of the rush hours, but there are also quite minor time differences. The place that is usually 30 minutes to the city centre can be 40 minutes just 10 minutes later. The maps thus highlight the rhythmic nature of public transport: it is not a smooth on-time provision of mobility. But the maps also show the spatial divergence with good connections along the routes but really limited for mainly low-density suburbanising housing in the North-Western part of the city. The opportunities of housing estates, where more than half of the city residents live (in Mustamäe, Lasnamäe and Väike-Õismäe) show divergence throughout the day but with travel times generally yet within 40 minutes. 

It is important to stress that the maps are based only on public transport timetables and the walk speed of pedestrians. This means that journey time depends only on the information contained in the timetable, and does not take into account the volume of traffic and traffic jams during peak hours. It is certainly a limitation but not that huge of an issue in Tallinn. On the one hand, schedules reflect lower speeds during rush hours, but on the other hand, the delays are less severe in this city of 430,000 inhabitants than in mega-cities. 

The basis for the whole analysis is the GTFS (General Transit Feed Specification) data which defines the common format of public transport data with associated geographical features. This format enables data sharing by public transport agencies and the creation of applications that use this data in an interoperable way. The GTFS data provides important information such as schedules, location of stops and all of the connections between them. For Tallinn, the GTFS data used in those analysis comes from Open Data Tallinn website (https://avaandmed.tallinn.ee/nimistu?id=33). Such data is available for many cities, thus making similar analysis doable across a number of places. 

There are yet some technical issues to consider with such analysis. Namely, preparing the proper road network layer, which is the basis for calculating walking and public transport travel, constitutes a challenge. When the network is incorrect, the whole analysis may give wrong results. Preparing the road layer for Tallinn entailed combining vector data from two different sources: Tallinn Geoportal layer (https://www.tallinn.ee/eng/geoportal/Spatial-data) and OpenStreetMap (OSM) created by the community of users based on information collected from a variety of sources, such as GPS or geographic databases. In both layers there are some deficiencies and errors. For instance, some areas were excluded from analysis because of missing road elements or wrong snapping. In the Tallinn Geoportal layer, some pedestrian paths were missing (e.g. in parks or in neighbourhoods between blocks of flats) whereas on the OSM layer some elements were not snapped to each other. Combining these data sources then gave more accurate information and after minor manual adjustments gave the results that hopefully most reflect reality. For Tallinn, all streets were assumed to be walkable. While this is not the case for each and every street, such assumption for this city is not that problematic. In some other cities, with more urban motorways, the picture can be quite different.

There are other similar maps in process for Brussels and Turku. If there are comments on this map, we are happy to hear from you. Please email Tauri Tuvikene (tauri.tuvikene@tlu.ee) or Agata Mielczarek directly (agatami@tlu.ee)