This project aims to predict the weekly sales of stores for a chain using regression techniques. Based on historical data, we analyzed various factors influencing sales, including holidays, temperature, the price of goods, Consumer Price Index (CPI), and the unemployment rate. Several regression models were constructed and evaluated to determine the most efficient.
- Project Overview
- Data Description
- Data Exploration
- Exploratory Data Analysis (EDA)
- Data Pre-processing
- Data Manipulation
- Feature Selection/Extraction
- Predictive Modelling
- Evaluation
- Conclusion
This project aims to predict the weekly sales of stores using regression techniques. We analyzed historical data covering the sales of 45 stores from 2010 to 2012, considering factors such as holidays, temperature, CPI, and unemployment rate.
The dataset includes:
- Number of stores: 45
- Time period: 2010 to 2012
- Variables: sales, holidays, temperature, CPI, unemployment rate, etc.
Initial exploration of the dataset, including basic statistics and data structure.
In-depth analysis to understand the data, including visualization and identification of patterns.
Steps taken to clean and preprocess the data before modeling.
Techniques applied to the data to make it suitable for modeling.
Methods used to select or extract features that are most relevant for the predictive models.
Different regression models used, such as:
- Linear Regression
- Polynomial Regression
- Ridge Regression
- Lasso Regression
Evaluation metrics and the performance of the models, with tables or charts if necessary.
Our project demonstrates the importance of data analysis and predictive modeling to understand the factors influencing sales and for strategies to improve them. The results obtained provide a a solid foundation for future studies and the implementation of concrete solutions The aim is to optimise the stores' sales performance.