TECHNICAL/FUNCTIONAL OVERVIEW

• Opt-in mobile phone app which GPS-tracks visitor locations real-time. Depending on visitor preference, app can either send fully anonymized data, or user-identifiable data to allow for such things as custom post-visit email containing visitor-museum-trail-exhibit-specific information of interest, entering visitor into a lottery (or every-N-ticket-sale gift) of free individual/family museum pass, etc.
• Statistics algorithm that converts incoming GPS data (from the subset of visitors that opt into app), to actual visitor count. For example, it might be determined that for each opt-in app user, one can expect 4 times that number of actual visitors. The connection between opt-in count to visitor count would be tuned, tweaked, and enhanced by comparing incoming app GPS head count with actual visitor count as recorded by human counters (potentially aided by live video) during "calibration" events.
• Backend server that receives incoming mobile phone GPS data in realtime, aggregating and segmenting same. The more segments available — based on how many visitor and environmental details the museum wishes to track, and how much data visitors are willing to supply — the more accurate the statistical conversion of incoming GPS data to actual head count will be. Examples include tracking such things as:
  • Time of day and season.
  • Weather.
  • What exhibit visitor is at, including whether inside or outside.
  • How visitor discovered app's existence.
  • If visitor agrees: Their age, interests, gender, nationality, etc.

  (Note: The time when statistic prediction is likely to be least accurate is when the number of visitors is low — a time when accurate tracking is least important.)

• App promotion campaign, which could include elements:
  • At Qatar airport.
  • On radio/tv.
  • On signs at exhibits themselves.
  • On museum brochures and/or tickets themselves.
• As many "dashboards" as desired (heatmaps etc.).

BENEFITS

• Overall, significant simplification of required hardware and software, as well as employing, in most cases, well established technologies that could be provided, maintained, and enhanced by many parties.
• Significant up-front development cost reduction. And dramatic cost reduction in ongoing maintenance — no array of outdoor networked CCTV cameras needing periodic service and replacement at all, just indoor backend server. All onsite hardware, and network connectivity, provided by visitors themselves.
• Accuracy rates that, to this author, seem likely to meet or exceed those possible/practical using CCTV-based methods.
• Scaling cost essentially 0.
• Ability to get a core system in place relatively quickly, and then enhance over time if and as needed, desired, and possible within budget constraints.
• Privacy of visitors, both real and perceived, much simpler. In other words, rather than a complex process to remove recognizable images from captured CCTV data, standard methods for anonymizing mobile phone data collection may be employed. This not only addresses the actual problem of privacy, but also the perceived one. In other words, if visitors see cameras everywhere photographing them, even if images properly handled, one also has to convince visitors of that fact (or hide the cameras, which introduces its own set of problems, both practical and public-relations-wise).

BONUSES

• For visitors wishing an enhanced experience:
  • Exhibit recommendations and enhanced exhibit information.
  • Personalized maps.
  • Direct feedback opportunities.
• Once the core infrastructure is in place (mobile app, backend server with database), enhancing the accuracy and depth of the correlation algorithms between collected app data and predicted actual visitor count should be straightforward. Enhancing how the app is promoted will similarly likely be straightforward.
  In consequence, there will be wonderful opportunities for local university students to contribute, and in so doing create connections, community, and unique real-world learning opportunities.
  Examples include:
  • Marketing and social science majors contribute to app promotion campaigns & offerings.
  • Statistics majors enhance the algorithm segmentation and prediction for improved accuracy.
  • Computer science/engineering majors enhance the functionality of the mobile app.