Inspired by web applications developed for San Francisco and New York, Sidewalk Labs’ newly launched Old Toronto (OldTO) website maps more than 30,000 (and growing!) historic city photographs from the City of Toronto Archives. A subsidiary of Google parent company Alphabet, Sidewalk Labs is working with the City of Toronto and Waterfront Toronto to develop a much-hyped “smart city” on the city’s industrial waterfront, though details of the plans remain relatively preliminary ahead of this week’s public meetings. For the time being, however, Sidewalk’s blog post has this to say:
The foundation of OldTO is the incredible collection of images at the city archives, which holds more than 1.7 million photographs dating back to 1856. More than 100,000 have been digitized and made available online, beginning with some of the most popular series by Arthur Goss (the city’s first official photographer) and William James (an early Canadian photojournalist). Other great series include subway construction photos from the Toronto Transit Commission and a set of 1972 street corners from the planning board.
To create Old Toronto, engineers at Sidewalk Labs used the titles of the photographs (such as “Yonge Street, looking north from Temperance Street”) to assign the image a latitude and longitude — a process known as “geocoding.” That makes it possible to map the old images onto Toronto’s current street network, creating a new way to explore the collection based on geography, and hopefully making the city archive’s valuable holdings more accessible to a wide audience.
Yonge Street, looking north from the CN railway, 1920.
Currently, clicking on a photo pulls up the information contained in the description of the archival holding, including (when available) the title, date, condition, and any copyright restrictions. Over time we plan to add even more images from both the city archives and other repositories. We’re also building an “aerial” button that will enable people to see how the city has changed from a bird’s-eye view (in addition to street view).
Geocoding is a great tool to help map a city, but it’s not perfect. City archive staff explained to us that some municipal addresses have changed over time as a result of street realignments or closures — even things like land reclamation along the waterfront pushing the shoreline further into Lake Ontario. One example is a 1928 image of a charming blockfront at “15 Bay Street”; today’s 15 Bay Street was nothing but green space at that time. (And as the archivists point out, it’s also possible that some addresses listed with the historic photographs are simply wrong.)
The engineering team at Sidewalk is currently exploring some additional features to make the tool more usable. One idea is using a computer to identify the various elements of an image — the awnings, streetcars, bicycles, etc. Doing so might enable people to search by keyword and pull up all the images with similar attributes, providing yet another point of comparison and exploration.
The map of images, via Sidewalk Labs.
We’re releasing the data for Old Toronto in the widely used JSON format, so software developers can easily use it to create new tools, visualizations, and interfaces. It’s a small example of our general philosophy that making data open increases its value to the broader community. We suspect there are many creative uses of this data that we haven’t thought of, and people will only come up with them when the data is easily available.
We’re also in the process of preparing OldTO for release as an open-source tool, so local urban-tech enthusiasts or civic hackers can build on top of it. (We’ll update this post when that release is ready.) With OldNYC, for example, two local engineers used the original tool to create a standalone mobile app. That way people can pull up a historic photo of whatever block they happen to be standing at as they make their way through the city.
As we think about the kinds of digital tools that help people develop, navigate, and maintain neighborhoods and cities, the ability to organize information geographically and by time comes up again and again as a critical requirement. Old Toronto relies on some of the same technologies that can support a future neighborhood, and as we continue to build new prototypes, we will use them to explore the digital infrastructure needed for more substantial applications.
A full copy of the blog post is available here.