- January 27, 2013
This was an ambitious data mining, mapping and statistics system for a startup called Tap-it. It was used to provide usage and location insight into the data coming from their app.
Tap-it developed an android app and delivered preloaded locked down tablets to rural trading stores across South Africa. The device and app gave traders a way to simplify their goods ordering cycle and gave wholesalers and distributors an effective way to communicate specials directly to the traders and check in and see that promotions were being set up and run properly.
There were over 1100 tablets in daily usage and the feedback from both traders and distributors was very positive but unfortunately the company struggled to find the right balance with their business plan and ceased operations in late 2013.
The backend of the android app was written in Ruby on Rails using a MongoDB database and the company who created it originally did not have bandwidth to create a statistics component.
I created a set of scripts that periodically dumped the important tables to csv files which were then automatically imported via feeds into a drupal site that I planned and built to mirror the key business objects of the rails app.
The site provided the following key features designed to assist support staff, salespeople and on the road representatives
A drag and drop customizable overview dashboard and map of key operations metrics and clustered trader distribution
Each map had the ability to be used fullscreen, had hashing to be able to deep link to specific locations and could be filtered further through a range of drop down menus
Maps of the store locations with different color pins to indicate their level of app usage
Proximity based searches to allow reps to build lists of traders and plan visits effectively
Distribution of key service providers displayed graphically on maps
A series of graphs and charts to allow the statistical info to be limited by date range and compared against each other.
The custom responsive theme was built using the twitter bootstrap framework, the site was built using an install profile approach and deployed and managed using Barracuda/Octopus Aegir.
Unfortunately when the company shut down the servers which were running the application, I had not taken any screenshots other than the one shown here. I will update this post with more screensnaps when I get a chance to set it up again locally.
- Data Vizualisation
- Mapping Drupal