Inspiration

We have seen that it has been a challenge to complete a simple trip to the grocery store. Many products are in shortage unexpectedly, even several weeks in the crises. Social Distancing can be a challenge in a full, tiny supermarket- and one would prefer going when fewer people are present. Normal busy hours don’t seem to apply anymore. So, how can we solve this issue of coordination? How can we empower consumers to make responsible decisions to where to buy what? How can we help risk groups to avoid crowds?.

Technology is my inspiration, especially artificial intelligence (AI). AI is being used in many sectors like health, automotive sector etc to solve problems which are unimaginable to be solved by humans. Having worked in the AI field for more than 4 years inspired me to adopt the use of machine learning in solving issues related to corona virus and prevent the further spread of viruses in the community.

What it does

Problem

To prevent the community spread of corona virus and flatten the curve, the entire world is in lock down. However, people still need to visit supermarkets and pharmacy stores in order to fulfil basic daily needs. Also, due to panic buying the entire supply chain has affected resulting in decrease of supply compared to the demand. This results in long queues in supermarkets and people need to visit many stores in search of very essential products which increase the chance of exposure to the virus. We want to help with better product coordination and more hygienic visits to the supermarket.

Solution

To overcome the above problem we are making use of crowd-sourcing and artificial intelligence and developing a mobile application named Crowd FREE which helps the people to get the latest update of the crowd in the nearby supermarkets or pharmacy stores and also get status of list of available essential products (Masks, toilet paper, sanitizers, chicken etc).

About the team

Mahesh Sulugodu Manjunatha : Team Leader:

I am a Master graduate with 4 years of experience in developing machine learning algorithms. As a team lead I was involved in pitching the idea and development of prototype mobile applications. Preparing the presentation and involved in the research phase of developing algorithms.

Somesh Sulugodu Manjunatha: Android Developer:

I am a bachelor graduate with 1.5 years of experience in developing android mobile applications. I was involved in developing the prototype of our application.

Ashwin Uchil: Agile Product Owner

I am an agile Product Owner with 10 years of experience in IT. As a team member I am helping the team to refine the idea and provide better customer experience.

How we built

The Crowd FREE mobile application provides two types of updates

  1. Crowd status update (example, percentage 50%,80%,30%) of the stores/shops.
  2. Update of products which are not available(example. masks, sanitizers) with partner shops/ when we have a critical amount of data.

How are we getting these updates?*

  • Crowd status update: We are using Bluetooth for getting the status of the crowd in the supermarket. When the user installs the Crowd FREE application, they should agree for continuous Bluetooth and location access. As soon as the user visits the supermarket the crowd free app scans all the nearby Bluetooth devices in the close proximity. Depending on the number of devices it updates the percentage of the crowd in that particular supermarket.

  • Update of products : In the initial stages either after visiting the store the application notifies the user to update the details in the application (Crowd source from the user ) or other option is to directly get stock details from the supermarket(Need to check with the stores and need to brainstorm how to approach the supermarket). This process continues until we collect sufficient data.

Use of Artificial Intelligence (Product status will be updated using machine learning algorithm)

This is the next stage of Crowd FREE app where AI is used for our prediction. Once the data is collected we will be using this data to train our neural networks (Artificial intelligence) algorithms. After attaining required accuracy (still in the research stage), this trained model will be deployed in the server. So that Mobile application will use these algorithms to predict real time status about the stores.

The solution’s impact to the crisis

Crowd FREE app helps all age group shoppers who are willing to visit the stores during the crisis.

  1. Old people: Since old people are more prone to get infected to the virus, With the Crowd FREE application they can check the status of the crowd and product availability. This prevents them from standing in the long queues and visiting multiple stores for the essential products. Also Family members can assist old people who doesn't know to use the application.

  2. People is need of sanitizers and masks: Since there is a shortage of masks and sanitizers Crowd free helps guiding the people to the nearby shops where it is availability and avoids people from visiting more than one pharmacy shops.

  3. Other people: With the Crowd FREE application we can prevent people from visiting multiple stores and spending more time in the queue in supermarkets, Which helps in preventing the further spread of the virus.

Use Cases:

When the Crowd Free application is installed, users should agree for continuous Bluetooth and location access.

  • Use Case 1: User intent to visit a supermarket and check the status (Before the visit).

When the user opens the application the on board screen will be the map view (Google maps API) where the user can search for nearby stores. And get the last updated status (these status are updated by the previous users who visited that particular supermarket).

  • Use Case 2: When the user visits the supermarket

The crowd free app scans all the nearby Bluetooth devices in close proximity. Depending on the number of devices it updates the percentage of the crowd in that particular supermarket. Also notifies the user to update the product status during their visit.

Current status of our application.

  • We have successfully developed a prototype of Crowd FREE application.

  • In the prototype application user can search for the near by supermarket with respect to his current location.

  • User can check the status.

  • User can update the status.

  • Research phase about the algorithm.

  • Prepared the future road plan.

  • Business model has been planned.

The necessities in order to continue the project.

  • Data collection: The machine learning model accuracy depends on the amount of data used for training. Since we are in the initial stages of development of mobile application.

  • Funding is need to hire more people to develop machine learning algorithm.

  • Funding needed to implement back ends servers through Amazon web services(AWS).

  • Store support to get the data.

  • Get market statics report.

The value of Crowd FREE application after the crisis

Even after a corona virus crisis, grocery shopping is an essential need for the people. According to statistics provided from Time use institute US grocery statistics. Below are some of the important information.

  1. Average time spent by a shopper in supermarkets is around 41 minutes.

  2. Average shoppers visit 1.5 times a week to supermarket.

  3. 40 % of shoppers visit more than one supermarket to get all the products.

  4. In an average grocery stores offers more than 42,200 items.

With these statics, below are some of the problems in this sector

  • People want to get in and get out of the supermarket as early as possible without getting stuck in the queues. But currently there is no option to get information, without visiting the supermarket.

  • Crowd management in supermarkets is a challenging job. Overcrowd increases the burden on the workers' staff.

  • Supermarkets are facing obstacles in understanding customer behaviour models which is very essential in supply chain management.

One stop solution is Crowd FREE app. Our idea of using Artificial intelligence helps in modelling the customer behaviour model and helps in crowd management. The accuracy of the machine learning model surpasses the human calculation which helps the supermarkets for data analytics.

Road plan

  1. Plan to release our mobile application prototype next week.
  2. Marketing the product and increasing the number of users.
  3. The amount of data collected is directly proportional to the number of users.
  4. Approaching the stakeholders (stores) for investment and pitching our product in angel investors.
  5. Expanding the tech team and start developing the AI algorithms.
  6. Expanding our development for IOS platform.
  7. Setting up back end servers in AWS.
  8. Deployment of our algorithms.

Business Model

Who are our customers ?

People are our customers. On an average 10,000 people visit grocery stores every day. No doubt there will be a huge customer base for our product.

Market size?

We can attract millions of users since grocery shopping is an essential need for the people and also it is increasing every day.

Is it Scalable ?

Initially we are confined to smaller region but easy to scale for millions of users. The only cost incurred is upgrading the back-end servers to support millions of users.

Cost incurred for our product

Since it is a mobile application our only investment is .

  1. Back-end Servers: Amazon provides less expensive back end servers which can be used for our deployment.

  2. Marketing.

  3. Development team: To develop machine learning algorithms can be done remotely.

Who are our investors?

  • Companies running supermarket chains: Since we are helping supermarket stores in tackling some of their obstacles like modelling the customer behaviour model and crowd management.

  • Private investors crowdfunding like angel investors.

Revenue generation:

  1. Subscription based business model user-centric approach: Since there will be million of users, If we are able to
    successfully collect just one euros per month, we can generate revenues around millions of euros as revenue.
  2. Subscription based business model for supermarkets: We can charge minimum subscriptions for each store of supermarket chains.
  3. Adding ads in our mobile application.

Viability

  1. Distribution channels: Android market store, later extended to IOS platforms.

  2. Competition: No mobile application which the proposed features.

  3. Legal requirement: Need to check with GDPR issues.

Built With

Share this project:

Updates