Emily Charles
  • Home
  • Work
  • About
  • Resume/CV
  • Contact

Smithsonian Visitor Center Analytics

While working as  a software developer for Boston Productions, Inc., I retrofitted an exhibit in the Smithsonian Visitor Center ("The Castle") with a comprehensive analytics system so that the institution could analyze and respond to visitor behavior. 

The interactive itself was a wayfinding station that features prominently in the lobby of the visitor center. Two portals allow visitors to select and read about any of the Smithsonian's institutions. From each of the portals, a special "Take me there" button can be selected, which animates a path from the visitor's location to the museum in question, as represented on a large map table between the two kiosks. 

My task was to implement a system of analytics such that daily activity logs could be collected and displayed visually for the Smithsonian staff. Data points included things like dwell time, popularity of certain destinations, and language selection. The details of the system are proprietary, but the mechanism consisted of three key components: data export, data transfer, and data visualization. 

​

Data Export

Building the data export process involved modifying currently-existing code to record important visitor actions, plus metadata that existed at the time of those actions. Given all of the requested data points, two distinct types of data emerged. I created two schemas, one for each data type, and added code to the existing base to export data points in a format that corresponded to the correct schema. 

Data Transfer

Due to the security restrictions imposed by the Smithsonian Institution, data transfer was a non-trivial task. The end result relies on several technologies working in a chain to ensure the complete and secure transfer of the analytics data. The transfer process is nearly 100% automated, and only requires staff interaction when generating new reports, again due to security restrictions.

Data Visualization

The visualization piece is handled by an enterprise-level application. It ingests the data produced by the map table and displays it in a variety of ways. My task was to configure a turnkey dashboard for the client that allowed slicing the data in several ways, including time granularity, aggregate vs. average, and language dependency. 
Picture
Photo credit: Boston Productions, Inc., 2015
Proudly powered by Weebly