Guideline for creating an OSM extract for HDX (using the automated osm2hdx tool)

Regions to create

The priority and status of countries to create as regions in the tool are given here.

Workflow

The ideal workflow is to

1. Create the datasets on HDX (using the guideline below) with the datasets set to 'private'.  This avoids having empty datasets visible on HDX while the first extracts are being created.
2. Once the export tool has finished, use the Admin UI to set the datasets to public. Do not set them to public using the HDX interface as this setting will be over-written the next time the export is executed.

Guidelines

1. Dataset name (title as seen on the HDX dataset page) should be HOTOSM [short version of country or region name]. 
   1. If the region being created is a country/territory in this list, then please use the short name found in the list.
   2. If the region being created is not found in the above list (i.e. it is a sub-national or multi-national region), then use a short but clear name in common use like "Lake Chad Basin" or "Sahel".
   3. An additional slug for the theme (roads, waterways, etc) and "(OpenStreetMap Export)" will be automatically appended.  
   4. Example: If you input HOTOSM Senegal, one of the output datasets will be HOTOSM Senegal Roads (OpenStreetMap Export)
2. Dataset prefix (URL slug) should be hotosm-[iso3 code or region name]

   1. Use the iso 3 code in this list or use the same name as for the dataset name.  Use underscores to separate words.

   2. An additional slug for the theme (roads, waterways, etc) will be automatically appended.  
   3. Example: if you input hotosm-fra, one of the output datasets will be at the url .../hotosm-fra-roads
3. Feature selection.  Unless you are sure, please use the default feature selection YAML that is automatically populated. 
4. Extra notes. Please do not populate this field unless you are sure.  It is normally left empty.
5. Locations.  Enter any locations that are fully or partially included in the boundary.  However, do not include adjacent areas that are part of the buffer around a selection region.  For example, if you are using the national boundary of Ethiopa and buffering it (see "Check Boxes" below), you would only enter "Ethiopia" in this field, not any of the neighboring countries.
6. Check boxes. Datasets should normally be public unless you want to test the output first and make it public later.  Extracts should always be subnational.  Buffer if using a national boundary (or other admin unit) to define the region.  This ensures that the desired region is completely covered.
7. Schedule.  Schedule should usually be set to weekly.  If you are creating a region for a new emergency, daily can be selected.  
8. Output file formats. All four output formats should be selected. 
9. Check the output box!  The output box should list 4 output datasets: buildings, roads, waterways, points of interest
10. Region definition.  
    1. You can draw your region in the tool.  If you need a freeform polygon, it's easier to draw it in geojson.io and save the file for import to the tool. 
    2. If you are using a shapefile as your source:
       1. make sure the data is in GCS WGS84
       2. multipart polygons are ok

       3. the ogr2ogr command to convert to geojson (and truncate the coordinate precision for minimum file size) is: `ogr2ogr -f "geojson" output_file_name.geojson input_file_name.shp -lco COORDINATE_PRECISION=5`

       4. If you are using an admin boundary of some kind to define the region, be sure to check the "buffer" check box.