Welcome to the Peer Lending Project, an analytics-based exploration of peer-lending investments for GreatYields. This comprehensive project is structured to answer key questions raised by Walter, the Chief Investment Officer at GreatYields, regarding the potential of incorporating peer-lending into the company's portfolios.
Peer lending, or peer-to-peer lending, involves lending money to individuals or small businesses through online platforms that match lenders with borrowers. The goal of this project is to leverage data science to gain insights into the investment potential of peer-lending notes, with a focus on data from SoftLending, a relatively new US peer-lending platform.
The project is divided into six stages, each with specific objectives:
Formally define project questions, identify potential issues, and analyze the case. Prepare a report and presentation outlining the questions to be addressed and potential challenges.
Ingest data, perform Extract, Transform, Load (ETL), and conduct Exploratory Data Analysis (EDA). Identify potential data issues and plan the project. Prepare a detailed report, presentation, and include the code used for data preparation.
Decide on the modeling approach, process the data, and justify decisions objectively. Prepare a report, presentation, and include the code used for data preparation and modeling.
Rely on an existing machine learning library, load data onto the selected model, and fine-tune the mining algorithm. Describe and justify the choice of model(s) and the required steps for training and testing. Prepare a detailed report and presentation.
Train, fine-tune, and test models. Evaluate model(s) performance statistically. Describe the training, tuning, testing process, and present the results. Prepare a comprehensive report and presentation.
Address business questions, provide answers, and investigate the business significance of findings. Suggest practical ways to apply results. Prepare a report and presentation summarizing the overall project, conclusions, and recommendations.
The project structure is organized into directories for each stage (a-f), and within each stage directory, you'll find subdirectories for presentations, working papers, and code. This organization facilitates easy navigation and collaboration.
Explore each stage's directory for detailed documentation, presentations, and code. For collaborations or significant changes, consider creating branches and following pull request workflows.