The metadata on the FSA-OWI corpus is incredibly rich, thanks to the meticulous work of the Library of Congress.  It includes a mixture of human-readable information, such as the photographer’s name, as well as fields that are a bit more… cryptic.

Exploring these MARC records, we became curious about a letter and number suffixed to the end of some of the call numbers. These numbers ranged from one through five, so attached to the call numbers was “-M1″, “-M2″, and so on. Looking more closer, we realized that some photos had an identical call number, with the only difference being this mysterious M number.  So, two photos could be LC-USF33-016166-M1 and then LC-USF33-016166-M2. We were excited about the possibilities these numbers, which seemed to imply a kind of sequence.  And they started make more sense once we explored the history of the FSA-OWI’s shooting method.

While we are used to the idea of a 36-exposure roll of camera film (at least those of us who remember shooting film!), the FSA-OWI photographers did not use this conventional standard. Instead, they took with them extremely long runs of film that were cut according to their needs. After taking their pictures, the film was sent to Washington, D.C. to be processed. During this procedure the long film rolls were cut into strips, most often of five exposures.

Since most of the photographs were digitized from the negatives,  the Library of Congress inadvertently provided essential information in the metadata that indicated the sequence of photographs within a given five-shot strip! Using the statistical program R and the code below, we were able to put the strips back together.


The result of this analysis was a strip of (usually) five photographs with their original order restored:

With the individual pictures put back into order — at least in sequences of five —  we can then examine the first and last picture in each strip. If we find an obvious match it means the end of one strip can be connected to the beginning of another. In many cases, we can then put the five-shot strips back into order, as part of the long film rolls that they originally came from.

We did precisely this with John Vachon’s trip from Kenosha, Wisconsin to Chicago, Illinois in July 1941.

The reconstruction of strips is not only a great visual way to represent the data, but also an exciting development for scholars.  We can now begin to track the journey of the photographers, as well as see their line of vision. In this case, we get a feeling for Vachon’s sense of humor.  Starting in Kenosha,  each photograph is of cheese, with the final shot in the sequence stating, “Stop Cheese.” This last image suggests both his playfulness — and his exhaustion!

Vachon becomes more serious as he transitions to Chicago: the first shot in that city is of a middle-aged man with his heard in his hand.  Out of millions of possibilities, Vachon chose this subject as his first image in the city; a critical stance as he captures the transition from abundance to poverty. The contrast between the Wisconsin cheese images and the man sitting on a Chicago street corner, presumably with access to few resources, augments the commentary by the photographer.

This is only the beginning of the type of analysis possible for this particular set of strips from Vachon’s trip through the Midwest, not to mention other strips and rolls from the collection. In future posts, we’ll look at how we can display photographs in their original sequence on a web page, to guide browsers in the footsteps of the original photographer and their  journey.

Posted in Uncategorized | Comments Off on Strips

Developing in the Wild

As a team, we have decided to develop in the wild.   We hope this will result in continuous feedback as we work to implement the project.  As a result, the website may be unavailable or change suddenly. We appreciate your patience and thoughts as work toward the official launch.  Any and all feedback is welcome.

– The Photogrammar Team

Posted in Uncategorized | Comments Off on Developing in the Wild

Beta Testing

Beta Testers,

Thank you very much for your help with testing Photogrammar.  We greatly appreciate your feedback. If you have any additional thoughts after today, please email Lauren Tilton.

Thank you again!

-The Photogrammar Team

Posted in Uncategorized | Comments Off on Beta Testing

CartoDB & Leaflet

I just finished getting the main photogrammar website linked to a table in a CartoDB database. This will eventually be the primary mapping mechanism for the site due to its scalability, nice balance of raster and vector data formats, and its quickly growing community of users. In doing this migration I have learned that my initial understanding of the relationship between Leaflet, CartoDB, and vizzuality was slightly incorrect.

Vizzuality is a smallish (~14 developers) firm based in Madrid and New York City. Following in the RedHat model, they provide fully open source tools under open source licenses but charge for support, dedicated servers, and custom solutions. These include CartoDB, CartoSet, and VisualRaster (see Vizzuality Tools). CloudMade is a similar, but distinct, company which concentrates on “Location Enabled Applications and Services”; there focus seems to be on mobile devices. One project they support is Leaflet, an “Open-Source JavaScript Library for Mobile-Friendly Interactive Maps”. While our focus is not on mobil apps, the leaflet library is known to be an excellent general purpose library for creating dynamic graphs. Due to the potential benefits from linking these two projects, Vizzuality has released an open source project on github for linking Leaflet and CartoDB together… In short, we have a lot of disjoint open source pieces from two companies coming together to produce a fast and robust mapping solution for the photogrammar website.

Posted in Uncategorized | Comments Off on CartoDB & Leaflet