Read This First

5/2018: The osm-editing-metrics is the spiritual successor to this repo. It uses the new osmium and its nodejs bindings for 🚀 performance.

12/2017: This project is a few years old and I haven't given it any attention. There are probably better tools out there to do quality analysis on OSM data now, especially at scale. Have a look at Mapbox's QA Tiles and some of the work done with them, for example by Jennings Anderson.

If you do find this project useful and are familiar with Osmium and its Javascript bindings, let me know if you want to take it over or become a maintainer!

Quality Metrics Suite

This is a growing set of OSMJS scripts that allow you to calculate quality matrics on OSM data. Currently it consists of:

• OSMQualityMetrics.js - various quality metrics, for normal OSM files
• UserStats.js - historical user metrics, for full history OSM files

Setup

You need the osmium framework for it to work, get osmium at https://github.com/joto/osmium and follow the install instructions given there. With Osmium set up, build OSMJS following the instructions at https://github.com/joto/osmium/tree/master/osmjs

Running the script

With OSMJS compiled, run one of the scripts:

    /path/to/osmjs -j OSMQualityMetrics.js -l array /path/to/data.osm

(It works equally well - perhaps even better - with a PBF input file, if you have PBF support in osmium)

The output on screen will look something like this:

    Running...
    parsing nodes...
    parsing ways...
    parsing relations...
    output and cleanup...
    total nodes / ways / relations: 51777 / 5040 / 28
    finished!
    Timings:
    total: 5002 ms
    ---------------------
    nodes: 4270ms
    ways: 617ms
    relations: 3ms
    overhead: 112ms

The scripts will generate output in the OUT_DIR specified in the script file.

For OSMQualityMetrics.js the output will be:

• ways.* : A shapefile containing all way geometries with version and timestamp attributes.
• metrostats.csv : The data metrics.
• tiger.csv : Breakdown of TIGER CFCC classes
• relations.csv : Breakdown of relation types
• userstats.csv : User breakdown

For UserStats.js the output will be:

• userstats.csv : Historical user breakdown

Sample output files are included in the example-output directory

Notes

• if you don't need the ways shapefile, you can set the OUTPUT_WAYS variable to false in the script. You can then also leave out the -l parameter when running the script and speed things up.
• The -l array option is best for large OSM files. If you're working with smaller, city-sized OSM data files, use -l sparsetable. Run osmjs -h
  for more info.
• The scripts will save its output files in the current working directory if no OUT_DIR is specified.

Timings

OSMQualityMetrics.js: On a Intel® Core™ i5-2410M CPU @ 2.30GHzx4 machine with 8GB of RAM running Ubuntu Oneiric, a 55MB bz2-compressed OSM XML file takes 103 seconds to process without way shapefile output. With way shapefile output using the sparsetable storage, the same file took 133 seconds to process.

Extras

The generated ways shapefiles will include version and timestamp attributes. You can use those to create interesting visualizations, like the ones shown in Martijn van Exel's talk at State Of The Map 2011. The Quantum GIS style file used to generate these images is included in the qgis folder.

What's Next

Easy

• Add way length statistics
• More attributes in ways output
• More specific stylings
• More specific statistics on relations

Harder

• More metrics on full history files