Authors

Inspiration

Shopping for groceries and supplies during quarantine measures is a critical task during this health pandemic. However this kind of activity can increase exposure to primary and secondary transmissions of COVID-19. The major precautions which need to be taken are:

  • Avoid overcrowding
  • Optimize travel outside the home.

People need their essential goods so why not make this activity safer, more informed and more efficient?

With the recent measure taken up by Coles (potentially other grocery chains in the very near future) to physically enforce customer numbers inside their stores, this issue has become realised. For the everyday person in their homes, there is a present need to provide an informed and centralised communication source - focusing on the customer density in grocery stores.

What it does

The central feature of our web application is a spatial heat map of grocery stores on a suburb-by-suburb basis, highlighting high and low customer density areas. Threshold values are used to clarify which suburbs are the most optimal for grocery shopping, updated every 15 minutes.

We also have the feature of push notifications of local stores that are currently experiencing low crowd density or stores which are deemed overcrowded (algorithm based or simply leveraging Big Data).

The web application itself is a comprehensive resource where customers can make a well educated decision whether to visit their chosen local grocery store (sent a personalized push notification alert). This is very important as it reinforces the practice of social distancing, monitored before you set off to the grocery store. When the heat map detects that now is the optimal time due to the fluctuating densities of customer headcount in the store, this alert is triggered.

To engage customer interaction, once they have checked the heat map by scrolling down the main webpage, their attention is now drawn towards an animated image. This image has been devised to remind customers to follow the safe COVID-19 health practices, highlighting the importance of protecting themselves and loved ones when grocery shopping. This image includes a hover over and a 'click me' button, where the customer will be directed straight to the Pre-Grocery and Post-Grocery COVID-19 Checklist. While observing the checklist they can easily navigate back to the heat map to check their local store.

Finally, to create a wholistic website, we have provided customers a live news update relating specifically to grocery shopping during this state of quarantine. With live updates such as Woolworths increasing their dedicated time spent disinfecting surfaces, the current Australian case count and breaking news of virus hotspots.

The technology-enhanced grocery store would feature methods to track customers entering and exiting the store and a form of verification for customer ending their grocery shop. Our hardware hack portion illustrates how such an automated customer throughput counter system could work.

How we built it

Website:

The user experience of this website (made with Wix) has been carefully considered such as the sophisticated but minimal design flow of the main page, with the menu bar at the top of the webpage, down to the 'Reach Out to Grocery & Me' contact form and a live chatbot offering the customer quick informative responses.

Hardware Proof of Concept:

An Arduino and LED array on breadboard demonstrates receipt tracking and customer throughput, generating the type of store data that could be channeled to such a spatial heat map. The green light represents people entering the store and the red light represents people leaving the store. The Finite State Machine and microcontroller timings 'simulate' random variability in customer numbers based on the rate of customer flow. Moreover, this kind of data is automatically logged at regular intervals.

Dataflow Pipeline:

The store processed data was then sent using the RS232 Protocol (Serial) through a COM port on a laptop and turned into a CSV file using a Python script. This data was then be ported into Alteryx/Tableau for more advanced data visualisation. The incoming data was formatted in a way to allow 20 suburbs to be represented in the subsequent heat map.

Heat Map:

The geospatial heat map was first generated using Alteryx Designer using an input CSV file of store ID, postcode and customer in store count at a regular time interval. Tableau was then used to shape this data into a proper heat distribution map with a legend for color-encoding warnings.

Challenges we ran into

The pipeline from Python to Alteryx was quite challenging as there was a need to have the data update in real time (ideally). In the end, our work around was to have the Arduino-store simulator data be configured into a CSV file. This file was then accessed and directly ported into the Alteryx pipeline to create the heat map. Implementation of the state machine with the various flags, timers and random logic was an interesting challenge as well.

Accomplishments that we are proud of

Marketing and Business:

The logo for Grocery & Me captures the essence of our aim. A moving animation of a busy crowd ties in well with the theme and our aim to assist people to buy groceries without a crowd - enhancing the grocery shop experience. A strong business and marketing model supports the tech development for our web application. We also made a neat survey with approx 40 responses within 24h to gauge responses regarding grocery shopping habits and attitude towards grocery shopping during these times.

Technological:

For the tech hardware hack, the randomisation of the LED displays and the visual red & green LED toggling of customer flow in and out of the store was successfully demonstrated. The Arduino - Python - CSV and Alteryx data pipelines were made effectively and cleanly to best generate and represent data related to our Grocery & Me solution.

What we learned

Through working on the Grocery & Me web application, our team learnt a great deal about how to best engage the Australian consumer. It also taught us to use our expertise for social and common good and aid everyone in their daily tasks during the quarantine.

What's next

With the advent of more strict quarantine measures in Australia, here are our next goals and steps:

  • A heat tracking map for online shopping based on web page queues could also be crated to extend the scope of the Grocery & Me application
  • Working out the costing and material requirements to implement this system in real life
  • Ideally we would like to partner with existing grocery chains to integrate this kind of data and metrics into their systems where possible to help as many people as possible continue to do their grocery shopping in a healthy way!

Acknowledgments

Try It out

Hackathons

Technologies

alteryx, arduino, c++, python

Devpost Software Identifier

257757