We were inspired by our own experience as introverted customers, and by the difficulties faced by a worker of the food service industry and member of our team ❤ Also, US restaurants waste up to 33 billion pounds of food each year (https://foodprint.org/issues/the-problem-of-food-waste/), and once the quarantine is over and people want to go out and see their friends, Nestor will be there to facilitate reservations and boost revenue for restaurants!
What it does
For restaurant owners, Nestor uses existing data to predict how average restaurants spend their resources every week - number of customers and number of each menu item ordered. Then, using data collected from the customer facing app, it trains a machine learning algorithm specific to each restaurant to predict the influx of customers and what their order will be. It allows for restaurants to know accurately how much food will be needed each week, thus reducing greatly the waste.
For restaurant customers, Nestor is the future of food services. It offers an intuitive and classy mobile app that lets customers reserve, choose what they want to eat, and pay in advance. This way, their meal experience is revolutionized : they come at the best time for them, their food is ready, and they can leave whenever they want since Nestor has already taken care of everything. Nestor makes the food service experience truly seamless.
How we built it
Our system is built with four main components: First, our AI prediction model is hosted on a python server and handles training and predicting for each restaurant. The data is taken from here https://www.kaggle.com/henslersoftware/19560-indian-takeaway-orders?select=restaurant-2-products-price.csv
Second, our REST API Node.js backend is hoted on Google Cloud and stores all important reservation and user data in a Google Cloud SQL Database.
Third, our business solution for restaurateurs is an accessible crossplatorm app built with the latest version of Angular that allows for owners to see upcoming reservations, monitor predictions for the future and customize their options (opening, closing, holidays, etc.) to allow full control on their brand
Fourth, our client facing app for customers is an intuitive mobile app built in Kotlin that allows customers to make reservations, easily discover new places to eat.
Challenges we ran into
It was for sure an ambitious project, and one of the main challenges was linking everything together in time and in a way that made sense. Having an artist on our team, we also put extra effort to make sure the UX was easy to use. Finally, we tried to build the best AI model in our limited timeframe.
Accomplishments that we're proud of
We managed to implement a functionning prototype for each part of the system. We also have over 60% accuracy on a model trained with real data. We are proud that for once, our app doesn't look like it was designed only by developpers ;)
What we learned
We learned how to implement a larger scale system that regroups multiple languages and libraries, and we learned how to work with an artist.
What's next for Nestor
There are so many potential features possible for Nestor. To summarize, we could personalize the customer experience and learn from their likings, propose dishes similar to what they like and restaurants similar to what their friends like. We could also add more features to empower restaurant owners, such as an internal customer rating system, an easier tip gestion, a floorplan to optimize employee's workloads and customer placements. There are many ways this app could develop, and we are excited to see where it can go.