This repository contains code and findings from a data science project aimed at predicting customer holiday bookings using machine learning techniques. The project involves exploring, preparing, training a predictive model, evaluating its performance, and summarizing findings in a PowerPoint presentation.
Getting Started Notebook: Jupyter Notebook for initial data exploration and basic dataset statistics. Python Scripts: Python files for data preparation, model training, and evaluation. PowerPoint Template: Template for the summary presentation. Resources Directory: Contains useful links and resources used throughout the project.
Explore Dataset: Understand the dataset columns and statistics using the provided Jupyter Notebook. Data Preparation: Processed data for modeling, created new features to enhance predictive power. Model Training: Used RandomForest algorithm to predict customer bookings and analyzed variable contributions. Model Evaluation: Conducted cross-validation, evaluated model performance, and visualized variable contributions. Summary Presentation: Compiled findings and insights in a PowerPoint slide using the provided template.
Getting Started: Follow the steps outlined in the Jupyter Notebook to understand the dataset and initial exploration. Data Preparation: Run the Python scripts to preprocess data and create new features as necessary. Model Training: Execute the model training script to build and analyze the RandomForest model. Model Evaluation: Run the evaluation script to conduct cross-validation and visualize variable contributions. Summary Presentation: Customize the PowerPoint template with your findings and conclusions.
https://www.kaggle.com/code/parulpandey/intrepreting-machine-learning-models
Python environment with required libraries (specified in requirements.txt ). Jupyter Notebook for data exploration and Python scripts for model training and evaluation.
Siddhant Ganvir Contact Information: sganvir204@gmail.com
The PowerPoint summarizing the findings is included as Customer Bookings