A time series forecasting project that predicts future sales trends for a retail business using Python. This project utilizes Prophet, Scikit-learn, and other essential data science tools for sales prediction and visualization.
This project focuses on building a sales forecasting model using historical retail data. It helps in identifying patterns, trends, and making informed business decisions.
- Forecast future sales using time series analysis.
- Identify sales trends and seasonality.
- Visualize and evaluate model performance.
- Time Series Forecasting
- Regression Modeling
- Trend & Seasonality Analysis
- Python
- Prophet
- Scikit-learn
- Pandas
- Matplotlib
- Historical retail sales data.
- Fields include item info, store info, and sales data.
- Frequency: Monthly (simulated from row count for modeling).
-
Data Preprocessing
- Handled missing values.
- Simulated date column for Prophet modeling.
- Encoded categorical variables for regression.
-
Prophet Model
- Trained for time series forecasting.
- Visualized trends, seasonality, and future predictions.
-
Regression Models
- Built additional models using Scikit-learn.
- Used
mean_absolute_errorandmean_squared_errorfor evaluation.
- MAE (Mean Absolute Error): 880.33
- RMSE (Root Mean Squared Error): 1093.54
- Visualizations show clear trends and seasonality patterns.
- Forecast plot with trends and uncertainty intervals.
- Component plots showing trend and seasonality.
- Sales over time.
- Fully working forecasting model.
- Evaluation metrics (MAE, RMSE).
- All visualizations embedded in the notebook.
This project is licensed under the Apache License 2.0 – see the LICENSE file for details.
- Facebook Prophet Team
- Scikit-learn Community
- Kaggle Retail Sales Datasets