Algocado

Inspiration

We were inspired by our backgrounds in data science and economics, and our experience and conversations with grocery stores about the large amount of fruits and vegetables thrown away every day. Through conversations yesterday with produce managers at 3 grocery stores, we learned that merchants throw out at least 6% of all fruits and vegetables they purchase due to spoilage.

What it does

Algocado increases profits and reduces the amount of waste created from the purchase and sale of perishable goods by advising merchants on dynamic pricing, longterm demand, and inventory management. Algocado uses machine learning methods and econometric models to predict the price and supply of avocados in the future, daily consumer elasticity, as well as the marginal cost of avocado inventory for a merchant. This information, combined with supplier data, yields supply and demand curves for avocados, from which we can derive a daily optimal price to maximize profit.

How we built it

We used React Js for our frontend. Our backend prediction algorithms are built using Python. Our data set is from the Hass Avocado growers association on avocado prices. The dataset includes data on daily avocado prices by region for consumers and stores.

Challenges we ran into

We ran into problems in finding large amounts of data on food sale prices and quantity sold. With more data we could make more advanced predictions.

Accomplishments that we're proud of

In less then a day we conducted customer interviews, mined data on food supply, and created a full stack analytics platform to optimally price grocery perishables. In doing so we also applied the economics and data analytics skills we've learned in the classroom to the real world.

What we learned

Through our conversations with merchants we learned about the inefficiencies of produce supply chain, storage, and sales which informed both our product and our interest in this area. Additionally, given we are looking at small segments of the market individually, we learned about what kinds of data we need to collect in order to predict optimal pricing, and how to best simplify multiple factors of demand.

What's next for Algocado

The principles and platform behind Algocado can be applied to any kind of produce or other perishable item with the right dataset. We plan to expand into numerous other industries that could benefit from dynamic pricing to improve profits, including movie theater and airline tickets sales. We have a vision for a smarter business that better captures and executes on sales data in order to minimize waste in all forms and maximize profits across all categories of items.

Additionally, we plan to integrate the dynamic pricing algorithm behind Algocado into our startup Snowball (meetsnowball.com). In doing so we will make Algocado's vision of reducing food waste a reality. The algorithms produced here will not only be used to price avocados, but all kinds of perishable goods including event tickets, restaurant item prices, and experiences. This integration will connect sophisticated dynamic pricing on the behalf of merchants to the consumer directly. Rolling out on a larger team buying platform has the potential to benefit merchants immensely while minimizing customer frustration over price discrimination, a common issue in dynamic pricing for goods with many alternative suppliers.

Built With

• econometrics
• firebase
• machine-learning
• python
• react-js