Disclaimer: We are not affiliated with the San Francisco Public Library in any way.
This site provides an alternative way of browsing the SFPL's incredible San Francisco Historical Photograph Collection. Its goal is to help you discover the history behind the places you see every day.
And, if you're lucky, maybe you'll even discover something about San Francisco's rich past that you never knew before!
The images all come from the San Francisco Public Library's San Francisco Historical Photograph Collection. They were culled from many sources, including the now-defunct San Francisco News-Call Bulletin.
The Library retains the copyright for many of these images. For details, please read their Permissions page and FAQ.
The creators of this site did not collect or digitize any of these images — credit for that massive undertaking belongs entirely to the Library.
The site was built by @danvdk and designed/built by @ravejk.
The creators of this site associated latitudes and longitudes to the images in the San Francisco Historical Photograph Collection. This process is known as geocoding. Doing this allows the images to be placed at points on a map, which enables new ways of exploring this collection.
The geocodes are based on two sources:
Several years ago, @danvdk searched for his cross-streets on the Library's San Francisco Historical Photograph Collection and found the photo on the right. The image was mislabeled — the intersection in the foreground is actually Waller and Fillmore, not Waller and Webster. Which meant that this photo from 1945 was taken from his roof!
He put together a now-and-then shot, but it always bothered him that the mislabeling of the image was so crucial to my finding it. This led to the idea of putting the images on a map.
And now, years later, we have that map!
The library's collection contains about 40,000 images. Many of these photographs have little geographic context (e.g. they're portraits) and cannot be located. In all, about 20,000 of the images could be placed on a map. We've geocoded about 65% of the possible images: 13,000.
If you're technically minded, here's a JSON file containing all the image descriptions, as well as geocodes for the records on the map (including the reason we thought they were at that location): records.js.zip (2MB download). If you improve on my geocoding or do something else interesting with the data, please share your results!