For TAMU Datathon

To run this, a folder named 'data' needs to be created with the appropriate data files added. I didn't include this file in the repo because I wasn't sure how public the data is.

In addition, an Gemini API key must be stored in a .env file in the structure GEMINI_API_KEY = "" for the chatbot to be initialized.

Mai KitShan was built to reorganize Ma Shan Yun's inventory such that it is easier to interpret and analyze. It provides several features:

1. Usage Comparison: Users will be able to see how the previous months usage of an ingredient compares to the actual usage of that ingredient in the current month.
2. Cost Prediction: The graph generated displays how much a customer is supposed to pay for a specific dish based on the ingredients used for it. A corresponding table highlights specific ingredients that cause customers to pay more or less for a dish.
3. Revenue Prediction: A bar or line graph displays the actual v. predicted revenues for each month, helping users with future budgeting or decisions.
4. Shipment v. Usage Analysis: Based on the shipping information of an ingredient, the user will be able to generate a graph that shows how much of a certain ingredient was shipped (in units of either pieces or counts) and compare it to how much of that ingredient was used in a given month. For better visualization, a table is created to translate the data from the graph. Another table is provided to suggest recommendations for future shipping orders, either suggesting to buy more or less of a certain item.
5. Usage/Shipment of an Ingredient: The user will have the ability to track how much of a certain ingredient is used v. shipped over time, helping identify seasonal trends for certain ingredients.
6. Bestsellers Analysis: A bar graph is generated to show the top selling menu items for each month or over time. Additional tables are present revealing the ingredients needed for each of the meals and what ingredient is the most frequently used among meals in the list of top sellers.