This project aims to predict store sales using the XGBoost Regressor model. The user interface (UI) is built using Streamlit, allowing users to interactively explore and make predictions based on input data.
The dataset used for training and prediction contains the following columns:
id: Unique identifier for each data entry.date: Date of the sales record.store_nbr: Store number.family: Product family/category.onpromotion: Representing whether the product is on promotion.
Example rows from the dataset:
| id | date | store_nbr | family | onpromotion |
|---|---|---|---|---|
| 1048544 | 2014-08-13 | 3 | BEAUTY | 0 |
| 1048545 | 2014-08-13 | 3 | BEVERAGES | 2 |
| 1048546 | 2014-08-13 | 3 | BOOKS | 0 |
| 1048547 | 2014-08-13 | 3 | BREAD/BAKERY | 0 |
| 1048548 | 2014-08-13 | 3 | CELEBRATION | 0 |
| 1048549 | 2014-08-13 | 3 | CLEANING | 3 |
| 1048550 | 2014-08-13 | 3 | DAIRY | 0 |
The XGBoost Regressor is used to predict store sales based on the provided features. The model is trained on historical data to capture patterns and relationships between the input features and sales.
from xgboost import XGBRegressor
# Code for model training...