Methodology & Data Sources

“Walk & the City” platform presents the first pan-European index of measuring the walkability level of major European Functional Urban Areas (F.U.A).

The European Walkability Map of our platform depicts an index assessed via the quantitative combination of eight (8) variables of available GIS-derived open geodata sources, in terms of the built environment characteristics, in order to spatially and qualitatively evaluate the levels of walkability or car-dependency in a user-defined area of any of the studied cities.

More specifically, the Walkability Index is composed by the following variables: the land-use mix, the population density, the “walkable” street network connectivity, the “walkable” street density, the pedestrian streets density, the access (400m) to public transport, the access (400m) to food stores and the slope.

The index has been internally processed and computed in a Geographic Information System environment and afterwards results for each urban area have been uploaded to the online Walkability Index Map.  Additionally, scores have been geo-visualized in a web-GIS module and the European 1km² Reference Grid (provided by the European Environment Agency as a free and open dataset) has been used for that purpose.

Below the 8 selected variables of the index are being analyzed more extensively:

1. The Land-Use Mix: An entropy index method is being used. The results range from 0 to 1, with 0 representing homogeneity and 1 representing heterogeneity. The values of this index are calculated for each cell of the grid via the following formula:

Where  i is the category of land-use, 

a is the proportion of the developed land area devoted to a specific land-use,

N is the number of land-use categories detected in each cell of the grid.

The data regarding the land-use categories are derived from the Urban Atlas European initiative of the E.U DG for Regional & Urban Policy and the DG for Enterprise and Industry with the support of the European Space Agency and the European Environment Agency. However, due to the fact that land use categories of this dataset are very general, the entropy score is based only in 2 land use categories, the residential and non-residential land use categories.

Regarding the residential land use category, it includes the following Urban Atlas land use types:

1. Continuous urban fabric (S.L>80%) (Code 11100),
2. Discontinuous dense urban fabric (S.L 50%-80%) (Code 11210),
3. Discontinuous medium density urban fabric (S.L 30%-50%0 (Code 11220),
4. Discontinuous low density urban fabric (S.L 10%-30%) (Code 11230) and
5. Discontinuous very low density urban fabric (S.L <10%) (Code 11240).

Pertaining to the Non-residential land use category, it includes the following Urban Atlas land use types:

1. Industrial, commercial, public, military and private units (Code 12100),
2. Railways and associated land (Code 12230),
3. Port Areas (Code 12300), Airports (Code 12400),
4. Green urban areas (Code 14100) and
5. Sports and leisure facilities (Code 14200).

2. Access to public transport (400m buffer zones): Access to public transport is fundamental for a walkable and accessible place. Thus, for each functional urban area we drew upon the relative data for public transport from openstreetmap.org (OSM) initiative. As a result, we created a 400m buffer zone polygon from each public transport point. Then, for each cell of the grid we calculated the public transport coverage percentage and the results ranged from 0 to 1.

3. Access to food stores (400m buffer zones): Walkable places should have in close proximity a specific coverage of some basic neighborhood facilities. To that end, we selected a group of food shops that could encourage people to do their daily errands on foot, if they are established close to their residence. We used point data from OSM (and particularly the following OSM tags: Supermarket, Bakery, Greengrocer, Butcher, Convenience, Kiosk, Mall) in order to create a 400m buffer zone polygon layer.  Then, for each cell of the grid we calculated the coverage percentage and the results ranged from 0 to 1.

4. “Walkable” Streets Network Connectivity: Street network data derived from the OSM dataset. We used the “Network Analyst” tool on our G.I.S environment and particularly the intersections density method (as a connectivity measure). First, we created the walkable street network dataset for each F.U.A and especially we selected only the streets with the following OSM tags: primary, secondary, tertiary, unclassified, residential, primary_link, secondary_link, tertiary_link, living_street_pedestrian, footway and path.  In essence, we excluded highways/motorways as they are being considered as not walkable segments of the network. Moreover, for each cell of the grid the number of intersections with more than 3 links has been calculated. Then, we standardized the values of intersections density measure with z-scores values and afterwards we normalized them on the scale from 0 to 1 in order for the results to be comparable with them from the other variables of the index.

5. “Walkable” Streets Density: Areas which belong in suburbia or have more rural characteristics tend to have less built-up public spaces (e.g streets). On the other hand, walkable places are dense with high density of walkable streets and with a high number of footpaths next to the street for the pedestrians to choose for in their routes. Hence, similarly with the aforementioned variable’s method, we used the walkable streets network derived from the OSM dataset and we measured the length density of these street types in each cell of the grid. Finally, the values have been standardized with z-scores and afterwards we normalized them on the scale from 0 to 1.

6. Pedestrian Streets Density: Places with pedestrian streets are car-free areas and usually have a livable urban environment which make them to be considered frequently as walkers’ paradises. In this variable we used the pedestrian streets data derived from the OSM dataset and we calculated the length of pedestrian streets intersecting with each cell of the grid. The values have been standardized with z-scores and afterwards we normalized them on the scale from 0 to 1.

7. Population Density: Population or residential density is a crucial factor for walkability, on grounds that it separates in general the compact urban form from the rural or suburban, rather sprawled, space.  The Eurostat’s GEOSTAT 2011 population-grid dataset has been used for the population density variable. Particularly, this dataset contains the count of persons at their usual place of residence per km² for the reference year 2011 on the date of the Census and it is geocoded to the INSPIRE Grid_ETRS89-LAEA_1K, which is the 1 km² European Reference Grid of the European Environment Agency. The values of population density have been standardized with the z-scores and afterwards we normalized them on scale from 0 to 1.

8. Slope: Walkable places should be comfortable places to move around for all people either for elderly or people with disabilities. To that end, slope is a crucial factor for walkability. Consequently, we utilized the EU-DEM v.1.0 dataset and we calculated the slope in degree values for each cell of the grid. Specifically, cells with mean slope value under 5 degrees gained 100% of the slope variable score, as being in general easy walkable surfaces. Cells with mean slope values over 15 degrees assigned with 0% of the slope variable score, while they are not comfortable and easy walkable spaces for all. Cells with mean slope values from 6 to 14 degrees gain the score resulted from the following formula: (15-i)/(15-5), where i is the mean slope value of a cell.

Table 1. Data sources of the Walkability Index