The PharmaSalesPredictor is a comprehensive Jupyter notebook designed for analyzing and predicting pharmaceutical sales. Built with PySpark, this notebook employs data processing, feature engineering, and machine learning techniques to forecast sales trends based on historical data.
- Data Cleaning and Preprocessing
- Exploratory Data Analysis (EDA) on pharmaceutical sales data
- Feature Engineering for predictive modeling
- Implementation of Linear Regression for sales prediction
- Prediction on both aggregate and individual product levels
- Exporting prediction results for further analysis
To use this notebook, you must have PySpark installed in your environment. The notebook is primarily intended for Google Colab, but it can be adapted for other environments that support PySpark.
- Clone the repository:
git clone https://github.com/[YourUsername]/PharmaSalesPredictor.git- Navigate to the cloned directory:
cd PharmaSalesPredictor- Open the
PharmaSalesPredictor.ipynbnotebook in Jupyter or Google Colab.
- PySpark
- Pandas
- Matplotlib (optional, for extended data visualization)
The dataset used in the notebook should be in CSV format and contain historical sales data of pharmaceutical products. The data preprocessing steps are tailored to handle specific data formats as detailed in the notebook.
- Data loading and preprocessing
- Exploratory analysis
- Data transformation and feature extraction
- Model training and evaluation
- Sales prediction and output generation
- Jean Paul, from Hit the Code Labs
Contributions to the PharmaSalesPredictor are welcome. Please ensure to update tests as appropriate.