2. Introduction
2.1. What is it all about?
The goal of the Accessibility Calculator plugin is to assess transport accessibility at the resolution of a single building. The assessment is based on a precise estimation of the travel time between the Origin (O), and the Destination (D) buildings of a trip. The OD travel time depends on the transportation mode which can be a Public Transport (PT) or a Private Car. The results of accessibility computations are stored as a CSV table that presents the details of trips and presented visually as an accessibility map.
2.2. Accessibility Calculator accounts for every component of a trip
The PT or car trip consists of several components: A PT user walks from the O-building to the initial stop, waits for a PT vehicle (bus, tram, metro, other), rides to the transfer stop, waits for the next PT vehicle, and, after, possibly, more transfers, alights at the final stop and walks to the D-building. The car traveler takes a walk from the O-building to the parked car, drives to the destination, finds a parking place nearby, and then walks to the D-building.
2.3. The Algorithms of Accessibility Calculator
To estimate transit accessibility we employ the RAPTOR algorithm.
To estimate accessibility with the private car we employ the Dijkstra algorithm.
Both algorithms are modified to deal with real-world data on the transit and road networks and to assess accessibility to city buildings (and not the network nodes or PT stops).
2.4. The data necessary for accessibility computations
To use the Accessibility Calculator, you need three sets of data, all covering the region of your interest:
The layer of roads.
The layer of buildings that are represented by polygons or points.
The GTFS dataset of the PT lines and their schedule.
The Accessibility Calculator checks the consistency of these three datasets and translates them into three fast-access databases. The first one contains a topologically cleaned road network, the second is built for computing transit accessibility and the third is built for computing car accessibility.
Note
It is often convenient to use datasets and build databases that cover an area that is larger than the region of the current interest.
In this tutorial, we use GTFS datasets that cover the entire Israel and exploit them for computing accessibility maps for the Tel Aviv Metropolitan Area. The Israeli GTFS datasets are available from https://gtfs.pro/, and in our examples, we use GTFS datasets for June 2018 and June 2024. The OSM layers of roads and buildings for Israel are available from https://www.openstreetmap.org/. In this tutorial, we use the gis_osm_buildings_a_free layer of buildings (as polygons) and the gis_osm_roads_free layer layer of roads. We also supply a zip file of all these layers as an appendix to this tutorial, see section 11.
Note
The layer of roads that you download from the OSM site often contains significant topological errors and must be topologically cleaned before using them for computations. This cleaning is a part of the road database construction, see section 4.2.
2.5. The menu of Accessibility Calculator
Figure 1 presents the menu of the Accessibility Calculator.
The Service area maps assess accessibility of the specific set of locations.
The Region maps assess accessibility of all locations in the area.
In this introductory section, we briefly present the abilities of the Accessibility Calculator. Each of these abilities is then presented in detail in the consequent sections.
Figure 1. The menu of the Accessibility Calculator
2.6. Data Preprocessing and Constructing Databases
To fix the data inconsistencies and accelerate data access for accessibility computations, all three datasets - roads, buildings, and GTFS, are cleaned and translated into internal databases. You can build several databases for different versions of the infrastructure or transit network development, assess accessibility for each of them, and then compare the results. The databases are usually constructed for large areas, or even the entire country. We recommend working with the datasets that represent areas containing 0.5-1M buildings.
2.7. From-accessibility versus To-accessibility
From-accessibility is based on the travel time from each of the selected buildings to all other locations in the city (Figure 2, left). The typical application of from-accessibility is the assessment of the residents’ travel time to the locations of their possible employment. To-accessibility is based on the travel time to each of the selected buildings from all other locations in the city (Figure 2, right). The typical application of the to-accessibility is assessment of the residents’ travel time to shops and attractions in the city center.
Figure 2. From-accessibility (left) versus To-accessibility (right)
2.8. Accessibility of several facilities versus accessibility of all buildings in the region
From- and To-accessibility computations can be performed to assess the service area of several facilities located in the buildings or to assess accessibility for all buildings in the region of interest – region accessibility. In both cases, origins and destinations can be a separate layer of buildings or just selected buildings. In the latter case, the selected buildings will be stored as a layer as a part of the results. The service area consists of buildings that can be served by at least one of the facilities.
The “From-” service area of the set of facilities includes all buildings that can be reached in maximum travel time or faster, from at least one of the facilities in this set. If the building can be reached from several facilities, then the fastest trip is considered.
The “To-” service area of the set of facilities, includes all buildings from which at least one of the facilities can be reached in maximum travel time or faster. If several facilities can be reached, then the fastest trip considered.
The details of each leg of the fastest trips are stored as attributes of the served building and the thematic map presents the minimal total travel time from the set of facilities to the served building or from the served building to the set of facilities. Importantly, the service area of every facility is also stored, including travel times from each of the facilities to or from every accessible building. This output is not mapped but can be exploited for computing other measures of accessibility, like the area from where the residents can reach more than half of the facilities in the center of the city in a maximum travel time. The region accessibility represents aggregate measures of accessibility for each building in the region.
The default aggregate measure of the from-accessibility for the region’s buildings is the number of other buildings accessible from each of them. Other measures, like the number of shops, or jobs, accessible from the building can be calculated if the information on jobs in the building or the building’s use is available.
The default aggregate measure of the to-accessibility for the region’s building is the number of other buildings from which it is accessible. Other measures, like the total population that can reach the building, can be calculated if the information on the population at a building resolution is available.
The accessibility measures of a region are stored as attributes of the region’s buildings, by the time bins defined by the user, like the number of buildings accessible in 5, 10, etc. minutes.
2.9. Adjustment of the trip’s start or arrival time to the transit timetable
Modern transit users are aware of the time the bus or train arrives at the stop they plan to start from, or to the final stops of their trip, and plan their trips accordingly. To assess the accessibility for these informed users we modify the RAPTOR algorithm to account for the schedule-based trip’s start or finish. Schedule-defined accessibility can be chosen for each of the From/To and Location/Region regimes.
2.10. Car speed for accessibility computation
To compute car accessibility, one must know traffic speed along the route. In the current version of the plugin, the traffic speed is defined by the type of road - a highway, major city street, neighborhood secondary street, etc., and the hour of the day. The free flow traffic speeds Vp, by the road link types p, is given in the car_speed_by_link_type.csv table (Figure 3, left).
The effect of the hour of the day is reflected by the Congestion Delay Index (CDI) - a ratio of the average, for the hour of a day, speed to the free flow speed, and stored in the cdi_index.csv table (Figure 3, right). Both tables are stored in the system folder and can be edited by the user.
The speed Vp(t) on the link of a type p at the hour t is calculated as Vp(t) = Vp*CDIt.
| link type | speed (km/h) |
|---|---|
| busway | 18 |
| cycleway | 15 |
| footway | 3 |
| motorway_link | 40 |
| ... | ... |
| hour | cdi |
|---|---|
| 0 | 1.0 |
| ... | ... |
| 5 | 0.9 |
| 6 | 0.65 |
| ... | ... |
Figure 3. Free flow speeds by the link types (left) and the CDI index, by hours of the day (right)
2.11. Comparison of accessibility in different scenarios
Typically, the accessibility is computed for different scenarios of urban transportation development, and then the outputs for these scenarios are compared. Let the scenarios be S1 and S2 and accessibility for each of them is already computed. The Accessibility Calculator includes three options for comparison scenarios’ outputs:
- Relative accessibility:
The ratio R1,2= S1/S2 of the outputs of S1 and S2, by buildings.
- Accessibility difference:
The difference D1,2= S1– S2 of the outputs of S1 and S2, by buildings.
- Relative accessibility difference:
The relative difference RD1,2= (S1– S2)/S2 of the outputs of S1 and S2, by buildings.
Note that scenarios S1 and S2 must be comparable, i.e., in sets of facilities must be similar in the case of the service areas or the regions, in the case of region accessibility, must overlap. The Accessibility Calculator tests the comparability of the scenarios, and either reports incomparability or performs the comparison.
2.12. Visualization of accessibility computations
The Accessibility Calculator results are always presented as thematic maps (Figure 4). These maps are based on one of the five layers that are constructed at the stage of data preparation (see Section 4), and one of them must be chosen by the user for constructing the map:
The layer of the Voronoi polygons constructed based on the buildings’ centroids, or
One of the four layers of:term:H3 hexagons <H3>, with the distance between centers 100, 200, 400, and 800 meters.
Figure 4: Left to right – The 45-min transit service maps from the Gesher theater in the center of the Yafo area, Tel Aviv, presented with the Voronoi polygons, 100 and 200 m hexagons
The visualization options are considered in detail in Section 9.