Jason Griffey, a consultant formerly at the Berkman Center for Internet and Society, Harvard University, was also librarian at the University of Tennessee. The library staff wanted to measure the attention of library visitors: where people tended to go and how long they stayed there, which can be thought of as similar to Google Analytics for a building.
Traditional metrics did not work well for the planning they wanted to do.
Commercial systems exist that are used by big-box retail stores to drive purchase decisions, but they are not appropriate for spaces where one might have an expectation of privacy. So the library decided to build its own system, which was built on open source software.
Possible ways to measure where people go include putting sensors on doors, tracking cell phones (but this ignores people not carrying a device, particularly children and recent immigrants). So they decided to use a camera as a sensor to see how movement happens in a space. No pictures are taken. Sensors were tested in the reading room at the New York Public Library (NYPL). A webcam was used as a sensor inside a 3D printed case.
The software is entirely open and runs on the Linus platform. It tracks moving pixels and stores location and time data. No pictures taken—only data paths. Results are displayed as a “heatmap” of activity counts. No photos or videos of library users are saved, and data are recorded in 15 minute blocks.
The NYPL wanted to know whether people were using the reading rooms to read or whether they were just taking a picture and then leaving (which is what they did). Over about two weeks, there were 480 visitors to one area and 2,630 to another. Most of them took a picture. Most of them did not move into the main space. The raw data were put into a database for analysis of counts and times.
At the University of Oklahoma, a local museum wanted to know how people were moving through it and which exhibits attracted the most people. An overlay of the data was created to see the popular exhibits. The goal of this project is to give it to as many people as possible so that spaces can be analyzed and changes tested. For example, one could evaluate how people use the stacks in a library and whether they are necessary, or give data to funders to prove how busy the spaces are. This project is built by librarians for librarians; it was not built by a vendor. See measurethefuture.net/pubicbeta.