Celery

Celery is hard, but green decisions don't have to be! We give businesses the data needed to make profitable & sustainable food!

Comment

Note: this project won Second Place overall and Best Financial Hack by Capital One as well as received honorable mention from iTrade (Supply Chain Data), BlackRock (Stock Market Data), and Google Cloud at LA Hacks

Inspiration

As a group of young and hungry students, we were inspired to pursue a topic near and dear to our hearts: food. Whenever you go to a restaurant, how do you know this meal won't be your last? Restaurants work hard to make their meals profitable and sustainable, but often lack the manpower and capital to get it right every time.

We are strong believers in the power of data to organize and transform lives for the better. Our goal coming into the hackathon was to create something that could help anyone understand the data surrounding them, regardless of technological expertise.

What it does

Celery is a mobile application for restaurant owners which predicts next months’ price of common foods, like Celery, and gives you feedback on how sustainable that item is right now.

Our simple neural network is able to anticipate food pricing chances using historic data from 1947 to now. In addition, this same network can predict your customer trends using your own purchasing data. This is shown in the last point on our trend graphs, showing the user where the market is headed.

Buying a food can be worse for the environment in different months because of many reasons, (transportation emissions, fertilizer runoff, and much more!), so we provide a simple metric to make the decision easy.

How we built it

After a few hours of brainstorming and scribbling furiously, we split our team into 3 frontend, 2 backend, and got to work.

Challenges we ran into

The largest challenge was the large amount of data management, integrating different tools, and developing time series algorithms (which is much harder on real data!). Most of the challenges stemmed from our hope to make a versatile software that would be able to scale as we grew our applications.

Accomplishments that we're proud of

The simple joy of getting code to work after 15 hours of errors, is what we enjoyed the most. I’m most proud of how my team went above and beyond throughout this process. Most of us have slept <5 hours across the weekend, and their dedication and engagement in this project is just awesome.

What we learned

We learned so much in so little time, across app development, server logistics, data preprocessing, and a hundred other tiny and frustrating steps. We all feel like we leveled up in our own specialties, whether it was working in Flutter for the first time or attempting to build a smooth data pipeline. We also learned a lot in the workshops, not just about the what but the Why we code, and the culture that surrounds our field.

What's next for Celery

Our first task is to make an application that can be attractive to a restaurant owner right now, and that means making reliable features and expanding our data set. We are very interested in working with iTradeNetwork, as their iFresh service could combine with our sustainability metric to help inform the decisions of stores. In addition, their Order management solution (OMS) records a large number of purchases, which could improve our pricing prediction and allow us to add a feature helping Owners find the best deal available across food suppliers, in terms of price and sustainability.

Our short term goals are to contact and survey restaurants around the LA area, especially those with a clear interest in sustainability, to refine our service. Our goal is to start with a app and web offering to companies that have an existing data architecture. Then, we would seek investment to work full time on the company, and expand our online offerings. The longer term goal is to have a platform that could be deployed in any restaurant, and provide data managment and analytical services.

Celery’s applications for streamlined profit management and state of the art prediction models do not stop with restaurants. We believe there is a serious need for an affordable, reliable, and robust system for helping small businesses make smarter investments.

Built With

Share this project:

Updates