At the end of every day, there's always some food left unbought in supermarkets.
The food is wasted if it gets thrown away.
So it's important to use them or give them away before they turn bad.
There are people who would benefit from getting free food.
For shop-owners, it can be so much hustle to go over all the leftovers and then finding foodbanks at the end of every work day.
So, Waste Fighters come with an automated solution!
What it does
Waste Fighters is a mobile application for supermarkets to donate leftovers.
The supermarket owners just need to capture a picture of leftovers on the app and the system will assign food tags to different foods.
They are then redirected to a page showing the nearest foodbanks based on the current location & app's Distribution System.
And then, using Twilio messaging, the app sends a message to the nearest foodbank, letting them know what foods are available.
How we built it
React-Native for application, Google cloud vision for labeling captured image, firebase as storage, Google Maps Platform API for finding nearby food banks, and Twilio for alerting messages.
Challenges we ran into
Detecting whether something is food using google cloud vision was one of the challenges. While Google Cloud Vision is really good at labeling images, it labels everything, not just the food! It is then up to us to determine if an object is "food" or "not food".
Obviously, the best way to overcome this is to use some type of machine learning algorithm. But the number of tasks needed to be complete and the time limitation, we're not able to come up with a model that's accurate enough.
To work around it, we downloaded datasets from Kaggle containing recipes. We then parse all the ingredients and use them as the main guide to determine if something is food or not food.
Another challenge was using the Google Maps Platform API. Getting the current location from the user was a challenge as that required permission from the phone user, which then has to be checked in the source code. The permission requirements also differed based on platforms, which made the process more complicated.
Accomplishments that we're proud of
For 2 out of 3 developers in the team, it was our first time working with React Native. There was a big learning curve that comes with learning mobile development, but we all thought it was a great learning experience.
What we learned
- React Native
- Google Maps Platform API
- Twilio API
- Kaggle has really good datasets!
What's next for Waste Fighters
As mentioned, the next step is to integrate some type of ML so that we can more accurately detect if an object is of type food.
We also want to add a chat feature or communication channel between the supermarkets and the foodbanks so the foodbanks can let supermarket knows in advanced what food they want.
All in all, we all believe Waste Fighters have a lot of potentials in helping both businesses and food banks. We are very excited to explore all these possibilities and see what Waste Fighters can offer!