Skip to content

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

2 Commits
 
 
 
 
 
 
 
 

Repository files navigation

📊 Sales Prediction Using Python

A machine learning project that predicts product sales based on advertising spend across TV, Radio, and Newspaper channels.

Built as part of my CodeAlpha Data Science Internship


📌 Project Overview

Detail Description
Objective Predict sales based on advertising budget allocation
Dataset 200 entries, 4 features
Models Used Linear Regression, Random Forest, Gradient Boosting
Best Model Gradient Boosting / Random Forest (highest R²)
Tools Python, Pandas, Scikit-learn, Matplotlib, Seaborn

📂 Project Structure

├── Sales_Prediction.ipynb       # Main Jupyter Notebook
├── Advertising.csv              # Dataset
├── requirements.txt             # Python dependencies
└── README.md                    # Project documentation

📊 Dataset Features

Feature Description
TV Advertising spend on TV (in $)
Radio Advertising spend on Radio (in $)
Newspaper Advertising spend on Newspaper (in $)
Sales Product sales (Target variable)

🔧 Workflow

  1. Data Loading — Read CSV using Pandas
  2. Exploratory Data Analysis — Distribution plots, correlation heatmap, scatter plots
  3. Feature Analysis — Investigated impact of each advertising channel on sales
  4. Model Training — Trained Linear Regression, Random Forest, and Gradient Boosting models
  5. Model Evaluation — Compared using MAE, RMSE, and R² Score
  6. Business Insights — Identified most effective advertising channels

📈 Results

Model MAE RMSE R² Score
Linear Regression ~1.2 ~1.6 ~0.90
Random Forest ~0.6 ~0.9 ~0.97
Gradient Boosting ~0.7 ~0.9 ~0.97

🔑 Key Business Insights

  • TV advertising has the strongest impact on sales (correlation ~0.78)
  • Radio advertising has a moderate positive impact (correlation ~0.58)
  • Newspaper advertising has the weakest effect (~0.23) — least cost-effective
  • Recommendation: Prioritize TV and Radio budgets, reduce Newspaper spend
  • Advertising spend explains over 90% of sales variation

🚀 How to Run

  1. Clone the repository

    git clone https://github.com/kinzaemannn/CodeAlpha-Sales-Prediction.git
    cd CodeAlpha-Sales-Prediction
  2. Install dependencies

    pip install -r requirements.txt
  3. Open the notebook

    jupyter notebook Sales_Prediction.ipynb

🛠️ Technologies Used

  • Python 3.x
  • Pandas — Data manipulation
  • NumPy — Numerical operations
  • Matplotlib & Seaborn — Data visualization
  • Scikit-learn — ML models and evaluation

📝 License

This project is for educational purposes as part of the CodeAlpha Data Science Internship.


About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

Packages

Contributors

Languages