Local Pitney Bowes Demographics
Mapping with Weather
An entry for the DevPost GE Intelligent World Hackathon
What it does
The Retail Traffic Insight application is a Predix cloud-based app designed to provide a real-time overview of vehicle and pedestrian traffic at individual retail stores.
The sample data represents a fictional, national big-box chain. Using intelligent parking lot sensors (vehicles in & out) and intelligent building sensors (people in & out), the flow of customers throughout the property can be monitored in real-time. By employing interior mobile sensors (possibly embedded in shopping carts) the fine-grained movement of customers throughout the store interior can be tracked.
Dwell time (how long a customer lingers in a certain area) can be very useful to determine customer interest in various classes of products (e.g. electronics, housewares, etc) as well as to measure the effectiveness of sales initiative in attracting customer interest.
Included visualization techniques include mapping at the nationa, regional and individual store level. Overlays for current traffic and weather can help detemine why traffic may be up or down at a given store.
Mapping at the store property level shows real-time activity in both the parking lot and individual store department level.
Finally, Pitney Bowes location-based demographic data in employed to chart population characteristics surrounding an individual store. This can be useful for planning sales initiatives and stocking levels.
How I built it
Using Predix CloudFoundry technology and Angular JS. Prefix provides a variety of services, e.g. Postgres, User Authentication (OAuth), Parking and Pedestrian Intelligent Environment sensors (All sensor info is simulated).
I used several standard Angular modules (e.g. Fabric.js, Google Maps) as well as Predix-provided UI components (e.g. data tables, alerts, etc.)
Challenges I ran into
Prefix takes an industry standard approach to development but there is a learning curve to be able to configure and use its services and APIs. Also you need to know CloudFoundry (which I did not).
What I learned
A lot about Predix!
What's next for Retail Traffic Insight
If there is any interest the next steps would be to replace the sensor data with some real-world sensor data at a pilot site. Expanded analytics using historical data could also be implemented.
User: iw_user_1 Password: IW_pass_1