Using mobile app usage data to inform the planning of urban infrastructure in public transport systems
As a technical consultant to the municipality of Patras (Greece), I have been involved in the planning and deployment of a public transport information system, including vehicle telematics, intelligent signage and public-facing mobile and web apps. My role in the project was to build the project specifications and monitor the project progress on behalf of the city.
Due to budget constraints, the project had enough funding to install 30 intelligent bus stop signs (in a total of over 1000 bus stops in the network). Some difficult decisions thus had to be made on where to install the 30 intelligent signs. To make these decisions, the selection process was facilitated by manually recording the number of passengers in several city centre stops, and by making some strategic choices (i.e. transport hubs, hospitals etc). We envisaged that the rest of the city would be covered by the mobile and web apps, which were to be developed as part of the project. The project specification dictated also that mobile and web apps should incorporate analytics mechanisms, to collect usage data.
After launching the mobile apps (on Android, iOS and Windows Phone), I examined the usage data collected by the incorporated mobile app analytics and made a quick mash-up that shows usage for the period of 27/11/2015 and 27/12/2015. The mashup shows the bus stops for which users have requested information at least once (represented by white dots), the locations where intelligent signage has been installed (red flags) and a heatmap, showing the frequency of bus arrival information requests by mobile users at each bus stop.
The visualisation confirms that, in most cases, the intelligent signs have been indeed installed at locations where a great demand for information is exhibited. It also highlights locations where information requirements may dictate a further need for installing additional intelligent signage in the future. Unfortunately, the intelligent signs were installed prior to the mobile app launches, hence we cannot monitor the effect of installing intelligent signs on the usage of the mobile apps, but this can be tracked in the future, when new signs are installed.
Other useful insights include the number of users, which for the given period amounted to approx. 3000 Android and only 200 iOS users and 200 WP users, as well as the number of sessions, amounting to an average of approx. 10 per Android and WP user compared to just 3 for iOS users. These data are perhaps indicative of social divides between user groups (iOS devices are prohibitively expensive to own for most Greeks). Further evidence of this divide is apparent through the fact that although the bus network covers most suburbs of Patras (from Tsoukaleika to the West to Aghios Vassilios to the East), the western (and least affluent) suburbs display a great demand for information, while there is no demand at all from the eastern (most affluent) suburbs (though it must be said that the eastern suburbs are also served by a popular suburban train route that extends from the city centre to Aghios Vassilios).
You can view the visualisation at http://18.104.22.168/patrasbus