In recent years, a handful of projects have used online map programs such as Google Maps to provide access to historic images of San Francisco. These programs allow website creators or users to “geotag” digital versions of historic images – in other words, to situate the digital images on an online map. Some of these projects are examples of Web 2.0, meaning that users are able to make contributions by uploading images of their own, or add tags or comments to others’ photos. By considering some of these projects together, it is possible to compare their features. I’ll discuss five projects here, and then raise three issues related to this kind of work.
One project is HistoGrafica. Users can upload images, as well as a free-text description of the image and its source. Users can also contribute to others’ photos by rating the image, adding it to a list of their favorites, recommending it to other users, or suggesting that the location assigned to the image may be incorrect. Images can be searched by location on the map or keywords that search the images’ descriptions, and search results can be limited to a range of years.
Another project is Sepia Town, by Jon Protas, Eric Warren and Eric Lehnartz. Like Histographica, users can upload photos, provide a free-text description, indicate the image’s location, and identify its source. Images can be searched by a location keyword, geographic coordinates, or using the map, but keyword searches do not search the descriptions of the images. Also, the search results cannot be limited by a range of years. Sepia Town does, however, offer a split-screen “Then and Now” feature allowing the user to see an historic photo on the left and a contemporary Google Street View image of the same location on the right. Of course, users can make mistakes in identifying the location of the images they’ve uploaded. Because the site doesn’t offer the option to tag or comment on others’ photos, users would have to contact the website owners to suggest that any mistakes had been made.
WhatWasThere is similar to Sepia Town in that users can search for images only by using the map or inputing keywords related to locations, and search results cannot be limited by a range of years. Whereas Sepia Town uses a split screen presentation to view both an historic image and a current shot of the same location on Google Street View, WhatWasThere allows users to superimpose the historic image directly over a shot from Google Street View and to adjust the historic image’s opacity using a slider bar. As with Sepia Town, in cases where an incorrect location is assigned to an image, there is not a means for other users to tag or comment on the image in order to suggest changes.
Sometimes a map-based set of San Francisco historic images is presented in a Web 1.0 format, instead of a Web 2.0 format. This means that only the creators of the site can make changes or new contributions to it. One example is Dan Vanderkam and Raven Keller‘s project, Old SF. Vanderkam and Keller have used images and their descriptions exclusively from one source – the San Francisco History Center‘s online collection of historic photographs. Users can search for images on the map and can limit the search results by a range of years. The Web 1.0 format means that if users have an interest in making their own contributions (such as providing additional images, or offering further information about images), they would contact the site creators rather than directly making the contributions on the site itself. A Web 1.0 format is easier to design than a Web 2.0 site, and it allows creators to have creative control over the site and its contents.
Vanderkam and Keller have done a tremendous amount of work in identifying the locations of images approximately 13,000 images from the collections of the San Francisco History Center. Building on that work, the San Francisco History Center has plans to further develop its own contributions to another Web 2.0 map project, Historypin. Historypin is perhaps the most ambitious of the current Web 2.0 mapping projects. In addition to features seen in the other projects discussed above, it also provides further options for user contributions – including upload of video and audio content.
A few issues arise with the emergence of map-based collections of historic images. One issue derives from the very fact that there is more than one project: researchers have to keep track of multiple sites in order to search for images they need. A similar challenge had been faced by researchers searching for digital versions of rare books, which archives had previously uploaded to a bewildering variety of locations. These locations included their own websites as well as third-party hosts such as Google Books, the Hathi Trust, and the Internet Archive. Recently, archives have been increasingly collaborating to aggregate their content into consolidated sites (such as the Medical Heritage Library project) both to reduce duplication of effort as well as to improve ease of research.
A second issue is copyright violations. Although the copyright of some historical images has expired, many site designers and contributors remain confused or uninformed about copyright law. Resources are available to clarify the copyright status of various works – one example is the excellent Copyright Chart by Peter Hirtle of Cornell University. Nevertheless, site users’ compliance with copyright law continues to be a challenge.
Finally, there are risks associated with the content of user contributions. Site creators cannot easily control or monitor contributions, and those contributions can include everything from innaccuate information (such as incorrectly identifying the locations of images on a map) to libelous statements or other inappropriate content. This year, the OCLC Social Metadata Working Group is publishing a report for archives, libraries and museums to address the challenges associated with hosting user-generated content on their websites. The first of the three parts of the report is already available for free online.
The projects considered here represent an ongoing attempt to improve access to historical images by making them available online, and by using crowdsourcing to geographically locate and describe those images. For more information about the projects, and to see how they continue to develop in both their features and content, please visit their websites: HistoGrafica, Sepia Town, WhatWasThere, Old SF, and Historypin.
2 responses to “Geotagging: Using Maps to Organize Historic Images”
Hi, thanks for this helpful overview… We’re trying to crowd source with an iPhone app the several thousand photos in the Shaping San Francisco collection at foundsf.org. (http://www.foundsf.org/index.php?title=The_FoundSF_Geotagger) I wish those guys who did the Library collection were interested in sharing their geotagging info, since a lot of our images are from the same collection. I’m also interested in how those of us who do this as a labor of love as opposed to a cash-generating operation can work with the materials that Historypin and SepiaTown are producing…
Thank you for your comment, and also for making mention of the Shaping San Francisco geotagging project. I think you’ve put your finger on one of the challenges: currently there are multiple geotagging projects, each with its own administrators, contributors, and arrangements for handling the intellectual property (such as geocoding) generated by these projects.
You also raised the question of whether collaborations among the aministrators of different projects might be feasible (such as administrators of one project allowing the use of their intellectual property by administrators of other projects). My sense is that exploring that possibility would require administrators of different projects to contact each other directly, and that willingness to collaborate might vary widely.
Regarding any possible collaboration between the Shaping San Francisco and the OldSF geotagging projects, I would want to note two things. First, Vanderkam and Keller had received prior permissions from the San Francisco History Center of the Main Public Library to use the images that appear on OldSF. Secondly, Vanderkam and Keller have made their geocoding of those images available for free in the form of a JSON file that can be found on the OldSF website. This means that their geocoding is available to anyone – including the administrators of the Shaping San Francisco geotagging project. In fact, as Dan Vanderkam says on that site, “If you improve on my geocoding or do something else interesting with the data, please share your results.” If you have further questions about the OldSF project, you might be be interested in contacting Vanderkam and Keller directly.