Measure the Future, an open source, open hardware project that enables libraries to collect and analyze data regarding how their physical space is being used, will soon deploy new sensors at libraries participating in its public beta launch, including the Meridian Library District (MLD), ID; the State University of New York, Potsdam; the New York Public Library; the University of Boston Law Library; and the Carnegie Library of Pittsburgh. Developer Clinton Freeman and project lead Jason Griffey (a 2009 LJ Mover & Shaker) recently changed the sensor hardware from Intel Edison to Raspberry Pi 3 single board computers, delaying the beta rollout, which was originally announced in late June. The switch will ultimately help ensure the project’s long-term sustainability on a more powerful, less expensive, and more widely available hardware platform. User observation studies are, by nature, time limited and labor intensive. And most new techniques for analyzing the movement and behavior of  users, such as Wi-Fi or beacon positioning systems, can raise privacy concerns, as well as excluding anyone who isn’t carrying a Wi-Fi- or Bluetooth-enabled device. By contrast, the computer vision sensors developed by Measure the Future use “blob detection” methods to collect completely anonymous data about patron behavior: counting people, determining where they are moving and how quickly, and how long they linger in a location. This information is  stored in a database; the “blob” photos are not retained. “I think there’s a lot to be learned from movement patterns, and how your spaces can be [made] better for the people who are using them,” Griffey told LJ. “The big difference between this and classic sociological observation is that it collects mounds of data that can be carved in different ways…. We have the ability to answer many of the same sorts of questions that sociological observation answers, like where do people stop? How long to they stop there? Are they stopping on the left side or right side of your new book display?” However, Griffey added, using computer vision sensors to collect all of this data, for an extended period of time, can reveal patterns that may not be detectable within the limitations of a human observer, but build up or change over time. “There’s a big difference between spending a week hand counting people and having a month, a semester, [or] a year” of data, he said. “It’s a much more thorough, and a much more robust set of data” than human observation could collect. And ultimately, Griffey plans for the data collected by the system to be mineable via web portal, enabling libraries to analyze patron behavior after the fact, comparing foot traffic patterns during different special events at a public library or analyzing finals week heat maps from different semesters at an academic library, for example. Griffey describes this as a “Google-Analytics-style dashboard for your library building.” “You [won’t] have to have the questions ready before you measure,” he said. “You can passively measure [using the sensors] and then choose to ask questions. What was the traffic like during that author event? How did it compare to the traffic at this other author event?... You can ask questions of the data that you might not think to ask before you have the data.”

Open source, open hardware

Originally funded with a $130,000 Knight News Challenge grant from the John S. and James L. Knight Foundation in 2015, Measure the Future has made its codebase available under a GNU general public license v3 via Github throughout its development. Detailed instructions for building and configuring the sensors—using a Raspberry Pi, Adafruit real time clock for Raspberry Pi, Logitech HD webcam, Micro SD card, Micro USB power supply, coin cell battery, and custom 3-D printable case—are available on the project’s official website at measurethefuture.net/build. At press time, the cost of the recommended parts in this list, minus the 3-D printable case, added up to about $104 per sensor. Although Griffey is offering to build and install Measure the Future systems for customers of his library technology consultancy, Evenly Distributed, he has long been an advocate for open source software and hardware, and said that there were many reasons why he will continue to keep project open as well. “The big [reason] is that it allows libraries that may not have the financial capacity to pay someone to install a system—if a library doesn’t want to contract with me or my company to put them in—it’s open. They can build it for a fairly low cost and tinker with it with fairly low risk,” he said. “Having the technology be available to the maximum number of people who may eventually want to use it is a big deal to me.” In addition, given the assurances that Griffey is making about patron privacy, he said it was important to enable others to analyze the system’s code. “Open code is more secure,” he said. “If someone wants to take a look at the code and verify that the system isn’t actually saving photographs, or wants to verify that the data is being saved in such a way that there aren’t unique identifiers…they can do that. You don’t have to take our word for it.”

Measuring potential

Meridian has been working with Griffey since the early stages of the project, and will be participating in the beta test as well. “Our role really was giving them feedback—being a space where we could say ‘this works, this doesn’t.’ We’re very excited about the next round,” MLD Director Gretchen Caserotti told LJ. “There’s been a number of times recently when we’ve been [thinking] ‘I can’t wait to get this data, because…we’re going through strategic planning, and we’re entering into a renovation planning stage.” MLD’s main library is 20 years old. While many minor changes have been made within the library over time, data about traffic patterns and how different areas of the library are used will provide important insights as MLD plans for a more significant renovation, she said. “Looking at high traffic [or] low traffic areas, where things are concentrated, [usage and traffic during different] times of day—all of that data would be useful for thinking about where we want to move things around.” Griffey agreed that renovation or rearrangement planning is one of the clearest use cases for these systems. But, he pointed out that libraries, museums, and other cultural institutions can also regularly use this data to complement other information about current use of their facilities and services. “I’m trying very hard to build a tool that gives libraries important information that we’re currently ignoring,” Griffey said, adding that he started the project because he “felt very strongly that libraries needed a new metric to express their worth to funders…. The traditional ways of justifying your budget—circulation numbers, through gate counts—those were becoming, in my mind, less interesting and reliable, and not as reflective of the work that libraries are doing these days. And part of the goal is to give people better answers, better stories to tell when you’re in front of your funders.”