Data import process into OpenStreetMap

Posted by DannyAiquipa on 16 October 2018 in English (English). Last updated on 17 October 2018.

One of the contributions could be used to improve the quality of OpenStreetMap is by importing existing datasets. However, if you don’t pay close attention, you may end up doing bad imports. Here, we collected some suggestions to guide you to import existing dataset into OSM correctly.

1. Type of data import

According to documentation in OpenStreetMap, only these physical features can be imported into OSM. They are:

  • Point: locations that can be presented by a single dot on the map, for instance, bank, restaurant or post office, etc.
  • Lines: the feature can be presented as lines, for instance, roads, trails, bike lanes.
  • Polygons: can be defined as a closed line, for instance, buildings, farmlands.

In the following picture, you can see how all the physical features that show and can be rendered on a digital map.


2. Data format for OSM

The default dataset extension of OSM is “.osm”. OSM data files are traditionally distributed in an XML format, and indeed it is still structured very much like XML. You can see an example as the following figure.



Converting other dataset to OSM In Java OpenStreetMap editor (JOSM), you can only edit feature data in ‘.osm’ format. If your data is packed in other data formats, we list all the possible conversion tools in the following table that you can use to convert your dataset into ‘.osm’.


3. OSM data license and permission

Once the dataset is in the correct format for OSM, you may consider OSM data license or permission. There are two kinds of scenarios:

  • Scenario 1. You own the data you are going to import into OSM, so you should consider an open source or public domain license. The public domain means it will be free and open to being used by anyone for any purpose without restriction under any copyright law.
  • Scenario 2. If the dataset originally owned by a third party, you need to have the permission to import the data into OSM. To obtain a permit, you can find some letter templates here. For specific details of the permission of data ownership, you can go and view it on Open Data Commons Open Database License(ODbL). If the data owner request, it needs to be addressed in the contributions specifically.

4. OSM data importation plan

You need to create a plan for any OSM data importation you are going to take, and the plan can be outlined as:


For specific you can see the following examples:

Register import plan. Edit and add your import plan in wiki OpenStreetMap Page: OSM Catalogue imports.

5. Data importation and communication with OSM community

Someone else may be working on importing the same dataset you have. Your dataset may be relevant to a region but not to other parts of the world. Therefore, keeping the OSM community informed of your intention is important. Before you import your dataset into OSM, please:

  • Send the import plan to country mailing lists.
  • Subscribe to import mailing lists to hear what the community has to say.
  • Send messages to local OSM users. You can find local OSM mapper groups by:

    • User groups: shows a map of the user groups by their geolocations;
    • Who’s around me? will show users who are close to your geolocations.
  • Also, send an email to image

  • Wait for approval of the import plan, if the local community has been approved your import plan, go ahead with the import process.

6. Other recommendations we have for you

  • Update your mapping project constantly.
  • Try to accomplish the schedule of your import plan.
  • When you are importing dataset into OSM, we recommend you to start with a smaller dataset. You should split a large area into smaller areas too.
  • Using high-resolution imagery, e.g. ESRI World Imagery, DigitalGlobe Standard Imagery, DigitalGlobe Premium Imagery, or Bing Aerial Imagery, to evaluate your dataset are recommended, but be minded that the imagery can be out of date.
  • Ask local mappers to validate and update your data that imported into OSM.


Let me know if you have any questions.

Comment from woodpeck on 17 October 2018 at 06:33

The strength of OSM, compared to closed commercial map offerings, is its community. If OSM were just a mix of imported third-party data sets, the HEREs and Googles of this world would long have overtaken us because they have more engineers than we have.

Any import into OSM must be seen through these eyes: Does it help the community? Is there a group of local mappers who were only waiting for this import to now start improving it and caring for the data? If yes, then that import can help OSM become better.

If, however, the import is done not by locals but by people who can be thousands of miles away, not “with” the community but “instead of” a community, then it is just window dressing: Making the map look better on the surface, but there’s no community to maintain it. Sadly, most imports are of this second kind, often done by people thousands of miles away with a phony promise of “making OSM better” - but OSM does not become better by some engineer hitting “upload” on a hundred thousand building footprints.

The claim that “one of the most important contributions you can make toward OpenStreetMap is by importing existing datasets” is fundamentally wrong, and demonstrates a deep misunderstanding of what OSM is all about.

Comment from kresp0 on 17 October 2018 at 06:33

Thank you!

I’m missing a link to the Import Guidelines, and to the Licence/Waiver and Permission Templates. There are also [translations into Spanish for the Waiver and Permission templates] (

My suggestion is to contact the local community first, update the proposal in the wiki and then contact the imports list.


Comment from DannyAiquipa on 17 October 2018 at 14:15

Hi woodpeck! Yes, you are right, It was my mistake, I re-edited this part of the diary. Thanks for your comment!

Comment from ᚛ᚐᚋᚐᚅᚇᚐ᚜ 🏳️‍🌈 on 17 October 2018 at 14:34

You say “The dataset you are ready to import to OSM cannot be older than two years”, and I’m pretty sure that’s not true. I’ve imported a dataset that started life decades ago, and is a few years old now. “Data quality” is important, and often associated with age, but one still have to look at the data, even if it’s “new”.

A tricky thing to describe is the emailing to the imports@ list. It’s not controlled by an import committee who will decide and vote and then tell you their result, so it’s hard to describe what “approval” means in this case. And if one person emails back “Looks good 👍” a new user might think that’s “approval”. “Seek consensus on the import@ mailing list” is more accurate, but vague if you’re not used to that form of decision making.

Comment from smaprs on 20 October 2018 at 12:14

Hi, nice practical advices to be in a check list. I remember another one: when using other file formats as sources, such as shapefile, always convert Coordinate Reference System (CRS) to EPSG:4326 - WGS84.

Login to leave a comment