Indianapolis Indiana ESRI Address Import

From OpenStreetMap Wiki
Jump to navigation Jump to search


To greatly enhance the address information for the city of Indianapolis, Indiana. The core of the city is woefully under-addressed comparatively to the outer ring. More information about the distribution of addresses my diary entry. The status of this project as of 1/2023 is Completed.


The import should start in late 2022-early 2023 and last 2-4 weeks, depending on how quickly I make progress. I managed a similar sized chunk of work in the Milwaukee import in ~10 days.

  • Data processing was done by ESRI in September 2020.
  • Local mappers are being contacted for discussion.
  • Submitted to import alias for review (12/10/2022)
  • Consensus seems to be that this level of documentation is overkill for this class of "import".
  • Working on getting a tasking manager setup.
  • Completed addition of ~150,000 addresses as of Jan 2023.

Import Data


I will be importing from the National Address Database dataset that has been prepared by ESRI. My estimate is that there's about 40,000 addresses missing from the city.

Import Type

A one time import that will be completed in smaller uploads as I make progress through the city.

Data Preparation

Data has been prepared by ESRI as outlined here.

Import Process

Data will be imported with my import account watmildon_import.

Pulling the data into JOSM is relatively simple using the MapWithAI plugin. After that my preferred workflow is to use the JOSM conflation tool. Fortunately, the vast majority of buildings already have outlines in the county from prior work.

  1. Set up conflation for the area of interest. (working OSM data layer and address information layer)
  2. With the Bing imagery layer, review all nodes that have not been associated with a building. Delete all nodes in vacant parcels. Merge into the data layer all nodes for buildings that require multiple nodes for addressing (ex: duplexes).
  3. Work through the conflicts list going item by item. This is a good time to spot check import addresses against already existing information.
  4. Sort the conflation list by distance and start reviewing items. Typically anything conflated more than 10m away has a decent likelihood of needing review.
  5. Visually review the conflation block by block. A typical issue to look for is an address node mis-conflated into a garage or other out building next to the main building.
  6. Conflate each set of blocks after review.
  7. Switch to the data layer and run Validation
  8. Review validation warnings and errors. Correct when necessary. The MapWithAI nearby street validation is great at catching subtle issues with data preparation (incorrect abbreviation expansion etc)
  9. Upload with these tags:

comment=Addresses in Indianapolis, IN



source=National Address Database