Level 1 - Road Network
In this exercise, we’ll investigate the quality of OSM data relating to road transport infrastructure. Start by choosing any region in Germany in the ohsome Dashboard for which you want to run the analysis.
Now you need to work out which topics in the ohsome Dashboard cover OSM data for road infrastructure. Take a look through the Topic Definitions to find out. Relevant topics include, for example: Roads (all highways), Roads (cars), Bridges (all ways), Bridges (cars) and Bridges (count)
For most use cases, the predefined topics are sufficient, and their naming is clear. Still, when choosing a topic it’s important to understand which feature types - in our case, which road types - are selected by the given definition. If necessary, the definition then needs to be adapted to your own requirements.
Topic: Roads (cars)¶
First, let’s build a general understanding of the OSM tags used when mapping road transport infrastructure. The OSM Map Features wiki page is a good place to start. Map Features describes all the basic OSM feature types and which additional attributes are typically used.
Solution
OSM uses the key highway for roads. You can find more information here. A more precise selection of the values used for road traffic can be found here. The topic Roads (cars) covers motorways, federal roads, state and county roads, and residential roads, along with their associated slip roads and link roads.
Taginfo is a very helpful resource for seeing how OSM values are distributed across OSM keys. Taginfo gives a good overview of which OSM tags are actually used, and how often.
Solution
How the individual highway tags are distributed in OSM can be examined via Taginfo.
The list of the most common tag values in Taginfo shows that, for example, service roads
(highway=service) and tracks (highway=track) aren’t covered by our topic Roads (cars).
Topic: Bridges (cars)¶
Bridges are also relevant to road traffic. You can read up on how they’re mapped in OSM here. Here too, you can investigate more information about their distribution in OSM on Taginfo.
This time we’re interested in the additional attributes used to describe bridges in detail. This information on useful OSM tag combinations can also be found in the OSM wiki and on Taginfo.
Solution
Taginfo lists the useful combinations in order of frequency. Bridges (7.2 million features in total) are frequently described in OSM with these further tags:
layer: ~6.7 million featureshighway: ~6.5 million featuresname: ~2.5 million featuressurface: ~2.2 million features
Choosing quality indicators¶
Different indicators are available for different topics. You can find these either in the Topic Definitions, or in the topic catalogue in the Dashboard.
Solution
Roads (cars)
Mapping Saturation
Attribute Completeness
Currentness
User Activity
Road Thematic Accuracy
Bridges (cars)
Mapping Saturation
Attribute Completeness
Currentness
User Activity
Running the analysis in the ohsome Dashboard¶
First, calculate the Mapping Saturation and Currentness for the topic Roads (cars) for the region you chose at the start. In the second step, you can change the topic to Cycleway for the same region and use the same indicators.
Next, we’ll extend the analysis to Attribute Completeness. We want to look at the attributes street name and surface.
Solution
Roads and cycleways:
Results for roads in Heidelberg: high currentness and high completeness at the same time
Results for cycleways in Heidelberg: high currentness, but not yet fully mapped in OSM
Explanation: the OSM community initially focused on mapping roads, and only later increasingly added cycleways. This process continues to this day.
Street name and surface:
Results for roads in Heidelberg: 67% street name and 83% surface
Results for cycleways in Heidelberg: 46% street name and 99% surface
Explanation: as expected, most cycleways don’t have a name. At the same time, cycleways in Heidelberg are mapped almost perfectly with regard to their surface.