This project for learning data science course, welcome for any suggestion and advice
Vending Machine need to re-supply each POS when depleted but Warehouse and Logistics process need time to planning route and prepare stock that will leave a gap of availability for sale
So if we can predict when each POS deplete beforehand for Warehouse planning and Logistics routing team can planning and prepare to re-supply daily or any periods
From Drink Vending Machine Company
Sample transaction logs from 3 different vending machine between 2023-10-01 to 2023-10-31
Explore all data aspects and split each machine's transaction logs to observe characteristics
add weekday,
then save to datasets/compute folder
you can see more detail at 2.LinearRegression.ipynb
you can see more detail at 3.ARIMA.ipynb
you can see more detail at 4.RNN.ipynb
with LSTM model from 2 implementation
- Still use rough indicator (CUPs) instead of in each ingredient details from drinking menus eg. matcha is run out
- Lack of menu recipe and machine capacity for each ingredients
- Limit on Time series information to 1 month OCT if we can gather more so we can analysis trends and seasonal
- Can add surround data to add more features eg. Weather, Building populations and characteristics,
- Finish this Proof of concept and pack into Application to use internal with MLOps