How geethasagarb/lead_scoring_case_study works
Main Components
• Data Sources:
• Leads.csv – the primary dataset
• Leads Data Dictionary.xlsx – metadata about the dataset
• Data Preparation & Cleaning: Steps that handle missing values and data variations (e.g., replacing 'Select' values with NaN and dropping columns with excessive missing data).
• Exploratory Data Analysis (EDA): Processes that generate initial insights (loading data, computing statistics, visualizing data through graphs).
• Model Building & Evaluation:
• Data Splitting (training and testing sets)
• Feature Scaling (standardization)
• Model Training (using Logistic Regression and possibly other techniques)
• Evaluation (calculating Accuracy, Sensitivity, Specificity, etc.)
• Reporting and Documentation:
• Jupyter Notebook (Lead Scoring Case Study.ipynb) where analysis is executed
• Supporting documents (e.g., Assignment Subjective Questions.docx, Executive Summary PDF, Presentation.pdf)
Relationships and Interactions
• That the Jupyter Notebook is central, serving as the interactive environment that ties all steps together.
• That reports and presentations derive their content from the outcomes of the analysis process.
Architectural Patterns and Design Principles
• Pipeline Structure: Clearly show the sequential pipeline approach (data ingestion → processing → analysis → model building → reporting).
• Modularity: Indicate that each stage (preprocessing, analysis, modeling, and evaluation) is implemented as a separate module or process, making the project maintainable and extensible.
• Reproducibility: Emphasize the role of documentation (README.md, data dictionary, and notebooks) to ensure repeatability of results.
Relevant Technologies and Libraries
• Python Libraries:
• Pandas and NumPy for data manipulation
• Matplotlib and Seaborn for visualization
• Scikit-learn for machine learning
• Statsmodels for additional statistical modeling