- Rebecca Pham
- May Lacdao
- Elizabeth Salas Martinez
- Hanieh Babaee
- Ronald Clarke
Deploy: https://retail-ml.herokuapp.com/
For this project, we decided to build on our two previous projects (ETL Project and Project 2). We will be incorporating machine learning algorithms to:
- Predict Walmart's stock price
- Create a grocery recommendation system
- Forecast sales
Data: https://www.kaggle.com/aayushkandpal/walmart-inc-stock-data-19722020-latest
Facebook Prophet
- run stock price vs date
- run stock volume vs date
- run (stock price vs volume) vs date
- graph stock data, prediction (line, area graph)
Data: https://www.kaggle.com/psparks/instacart-market-basket-analysis
A. Grocery List Recommendation:
- Load csv file, build a clustering/classification model
- Webscrape/API google images, use chrome driver searching the word/food item and just taking the copy link address of the first search results, adding the csv file
- HTML, search a member number and the product suggestion will load
B. Other Product Recommendations:
- Additional Feature: “You May Also Like..."
- Using Surprise Algorithm provide 3 products a user may also like to purhcase base on their order history
Data: https://www.kaggle.com/naresh31/walmart-recruiting-store-sales-forecasting
A. Sales vs. SocioEconomic Target
- Regression on Sales vs SocioEconomic Variables (No Date)
- Regression to see if there’s relationship between sales and socioeconomic variables
- Graph against the variables to see if there’s any relationship
B. Facebook Prophet Sales vs Date
- If there’s no relationship above, you can use facebook prophet
- Graph a line/area graph with the sales and predictions
