Inspiration
Since the pandemic struck, local supermarkets have faced surges in demand whenever Covid-19 restrictions tighten, disrupting the supply of groceries to households.
What does our solution do?
Our 2-pronged approach aims to increase supermarket sales through identifying complementary products using association rules and recommending them to consumers. Additionally, by analyzing past sales data, we are able to forecast periods of high demand to prevent under-supplying of goods and lower inventory costs.
How did we obtain the model
We first did exploratory data analysis by using Simple Moving Average model and plotting graphs to uncover relationships between items sold, day, month and year. Next, we used Long-Short Term Memory Recurrent Neural Network (LSTM RNN) Model to forecast future sales to better anticipate supply shocks for the super market.
Problems faced
We faced problems running R on Google Collab and thus could not collaborate together to help each other on our codes. We also had difficulties running the LSTM RNN Model as some of our laptops keep crashing whenever we run it.
Things we are proud of
We managed to generate logical solutions for both consumers and the supermarket within this short amount of time. Every team member did their part to contribute actively to this project.
What did we learn?
We are more familiar with using association rules and creating neural network models in R.
What is next for Market Basket Analysis?
We hope to expand the scope of our project to other industries, such as ecommerce to benefit more consumers and companies.
Built With
- gganimate
- keras
- r
- tensorflow
- tseries
Log in or sign up for Devpost to join the conversation.