Making salads great again!

Optimisation of Salad Stop's business model in order to help improve their business decisions and profits.

1. Overview

• Descriptive Analytics: Visualisation and formulation of business problem

  1. Relationship between complicated menus and purchase probability

  2. Trends of time-dependent demand of ingredient

• Predictive Analytics: Data modelling and forecasting

  1. Predict purchase probabilities given Salad Stop's menu size

  2. Forecast future demand of ingredients

• Prescriptve Analytics: Optimisation

  1. Linear Optimisation - shorten menu choices provided by Salad Stop with customized constraints

  2. Stochastic Optimisation - optimise the amount of ingredients to import in the following day/week/month

• Proof of Concept solution

  • Demonstration video

  • Web application with a proper UI serving our optimisation models for end users

  • Install all relevant libraries before running commands

    cd poc
    flask run 
    

  • / home directory for linear model

  • /stochastic directory for stochastic model

2. Development using Git Branching

This is to prevent conflicts when we are pulling/pushing code. Finalised model/code will be merged into master later on.

git branch YOUR_NAME
git checkout YOUR_NAME
git add .
git commit -m "COMMIT MESSAGE"
git push origin YOUR_NAME

3. Dev Notes

1. Before developing/analysis, generate datasets first (make sure you have installed the appropriate python libraries)

   cd code
   python3 generate_datasets.py
   

2. NotImplementedError: Cannot convert a symbolic Tensor (lstm_2/strided_slice:0) to a numpy array

   • This is a result of conflicting numpy versions (RSOME requires the numpy version to be 1.16 where as the LSTM model requires numpy version to be 1.18)

   • Fix: Solved by modifying tensorflow/python/framework/ops.py as seen here