How nooranalytics/customer-analysis works
Project type and purpose
• This is a data-analysis/EDA project (not a full-stack app). It uses Jupyter (analysis.ipynb) to explore user and purchase CSV data for a Meditation App, with the goal of understanding customer demographics, conversion rates, and revenue patterns.
File structure
• Top-level:
• analysis.ipynb – main analysis notebook
• data/ – all raw CSV inputs
• data/ contains:
• user_demographics_v1.csv
• user_demographics_paywall.csv
• purchase_data_v1.csv
• AB_testing_exercise.csv
• daily_revenue.csv
• No backend/server code, no deployment scripts—purely data ingestion→analysis→visualization.
Extract additional insights from README
• Reads two primary CSVs (demographics + purchase), but extra files suggest A/B testing and revenue‐over‐time analyses additions.
• Dependencies: pandas, numpy, scipy, statsmodels, seaborn, matplotlib, scikit-learn, pingouin.
• Workflow is: clone→install libs→open notebook→run.
a. Main components
1) Data Sources
• Five CSV files in data/
2) Data Ingestion Layer
• pandas I/O reading CSVs
3) Data Processing & Transformation
• pandas/numpy operations, merging, cleaning
4) Statistical & ML Modules
• scipy/statsmodels for significance tests
• pingouin for A/B testing
• scikit-learn for any modeling
5) Visualization Layer
• matplotlib/seaborn charts
6) Presentation Layer
• Jupyter notebook interface
b. Relationships & interactions
• Ingestion → Processing/Transformation
• Processing → Statistical modules (branch)
• Processing → Visualization
• Statistical modules → Visualization
• All outputs loop back into analysis.ipynb
c. Architectural patterns / design principles
• Linear data pipeline: Data → ETL → Analysis → Visualization
• Single orchestrator (Jupyter notebook) drives workflow
• Modular use of libraries for separable concerns (I/O vs. stats vs. viz)
d. Key technologies / libraries
• pandas for DataFrame operations
• numpy for numerical arrays
• scipy/statsmodels/pingouin for inferential stats
• scikit-learn for predictive modeling
• seaborn/matplotlib for plots
• Jupyter Notebook as the interactive shell