A machine learning-based automation tool that intelligently predicts monthly sales targets for each brand based on historical targets and achievements. This system is ideal for organizations that set monthly targets per brand and evaluate performance yearly.
Sales planning often relies on historical performance data to define future targets. This project automates that process by leveraging machine learning — specifically a RandomForestRegressor — to predict future monthly targets using past achievements and trends.
The model uses historical data stored in a CSV file containing monthly targets and achievements per brand, and it outputs predicted targets for any given month or even for the entire year.
- 📊 Reads and processes historical brand performance data
- 🔁 Predicts target for a single month or the entire year (Jan–Dec)
- 🧠 Uses scikit-learn's Random Forest Regressor for prediction
- ⚙️ Handles feature engineering automatically
- 📁 Option to save or export predictions
- 🛠️ Easy to customize and retrain
Your input CSV file (brand_targets.csv) should contain the following columns:
| Brand | Month | Year | Target | Achievement |
|---|---|---|---|---|
| BrandA | Jan | 2023 | 1000 | 950 |
| BrandA | Feb | 2023 | 1200 | 1100 |
| ... | ... | ... | ... | ... |
⚠️ This version supports a single brand setup. Multi-brand support can be added later via one-hot encoding.
- Python 3.x
- Pandas
- scikit-learn
- Joblib
- Clone the repo:
git clone https://github.com/yourusername/brand-target-predictor.git cd brand-target-predictor - Install dependencies
- Place your dataset (brand_targets.csv) in the root directory.
- Train the model
- Predict a single month or Predict a full year (Jan-Dec)
- ✅ Multi-brand support
- 📈 Integration with real-time dashboards
- 📁 Exporting results to Excel/CSV
- 🧠 Model optimization using time-series forecasting
Adewale Abdulmuiz Akorede Email
⭐ If you found this useful, give it a star and share with your data and business intelligence community! LinkedIn
⚠️ Ready to collaborate with a Real time-series forcaster programmer