In March this year, the city of Montreal released nearly four thousand aerial photos of Montreal taken between 1947 and 1949, which form a complete survey of the whole island. This is an great and incredibly detailed resource. Unfortunately, the data was published as as a large set of image files without any geo-references.

So while this is a great resource for the community, it is incredibly difficult to use, and very hard to browse. In order to make it more useful, I stitched some of the images together, aligned with the modern satellite view of google maps, and made a browsable map to easily compare then and now (http://www.app.catbus.ca/1947satelliteview/app.html).

For this effort, I used python as the glue of the project. It allows me to script the generation of thousands of images, while at the same time juggle the math to project the imagery onto the modern map. I'm in the process of expanding this project to include the whole island.

In the talk I'll explain

• the data source, and how hard it is to access, i.e. the motivation for the project
• the process of combining and aligning imagery
• how large browsable/zoomable maps work by breaking the imagery into tiles
• the ongoing project to expand this project to cover the whole island, and the creation of open source tools that allow the creation of geo-rectified aerial tile sets from aerial photos
• a note on open data, and how making it accessible like this allows everybody to find their story in the data