open navigation menu close navigation menu

The Online Language Archive at the University of Chicago

A digital audio tape. Photographer: Mark Ramsay.

This archive contains recordings of 270 different langauges, with examples dating back to the 1890s. As sound technology has evolved, wax cylinders and phonographs have been transferred to magnetic tape and finally to digital audio, which is now searchable on the web.

The project presented a few challenges. First, the data was being managed in a FileMaker Pro database that had evolved substantially over time. I extracted the data from that system and transferred it into a graph database using linked data triples.

Second, the data itself is nuanced and rich, and as we worked to transform it to linked data the reality of the current data model often didn’t match the model of our target system. As we worked I wrote ad hoc scripts, and produced visualizations and reports to communicate about various data issues with stakeholders.

We worked under a very tight deadline to produce a web interface for the project. I project-managed that part of the project, organizing the work of a developer and designer to be able to launch the site on time and with the features our users needed.

Visualizing format relationships

A subgraph of format relationships in the system. Arrows point from original formats, like LPs, to descendant formats like digital audio files. In this diagram several LPs were converted to a single 1/4" tape, and several digital files were made from that tape. Visualizations like this help when debugging data issues.