This past fall, Elliot Plack, who then worked for Baltimore County GIS and was appointed to the Maryland Open Data Council by Governor O’Malley, got in touch with me about restarting the import, after finishing an import of similar data for the County. After meeting with Jim Garcia from MOIT, he was able to secure the permissions that we needed to proceed with the import. Additionally, he was able to get address point data, which is much superior to and easier to use than the parcel data I was originally going to extract addresses from.

With the address data in hand, I worked through the remaining technical aspect of the import in January and February with help from Elliot.² Besides simplifying the building footprints,³ footprints intersecting existing footprints needed to be separated out, and addresses needed to be added in. Where buildings didn’t intersect existing footprints and only contained one address point, the address was merged into the building. Addresses that were contained in intersecting buildings were also separated out. Address data also needed to be translated into OSM tags. We decided to automate the import as much as possible, with non-intersecting buildings and addresses to be uploaded via script. Conflicting data would then be manually conflated and uploaded.

About a month ago, with the preparation done, Elliot sent an email to the OSM Imports and US Imports lists announcing the restarted import. After receiving no objections, I executed the automated portion of the import a week and a half later. A job was then set up on the openstreetmap.us tasking server for the manual portion of the import,⁴ which is currently underway. Elliot and I as well as a couple other local mappers have been working through this portion of the import and should hopefully be finished in the next month or two, which will conclude the import.

1. http://wiki.openstreetmap.org/wiki/Baltimore_Buildings_Import ↩

2. The code used for the import is on GitHub. ↩

3. This is an interesting problem, since the topology of adjacent row houses needs to be accounted for. ↩

4. http://tasks.openstreetmap.us/job/59 ↩

This time, I designed an aluminum rig to hold the DSLR and suspended it between the railings using parachute cord. After stringing the cord across the atrium and pulling it taut with a trucker’s hitch, I attached the camera and pulled the rig to the center of the atrium using fishing line. Once the camera stopped swinging, I took around a dozen exposure-bracketed sets of photos via USB tethering and then took everything down. Finally, I used Hugin to correct the lens distortion and small misalignments for three bracketed sets, masking out people who were moving in each set; Enblend and Enfuse were used to create the final image, which was then cropped.

A vector-based PDF version suitable for high-resolution printing is available here. The map data is copyright OpenStreetMap contributors under the ODbL. The map itself is CC BY-SA Stamen Design (and me). Here is a photo of the map on my wall. Feel free to get the PDF printed.

Flickr