How eppursimuove9/bank-loan-analysis works
Project type and purpose
• This is a data-analysis/BI pipeline focused on bank loan portfolio performance. It uses raw loan data (CSV), SQL for extraction and transformation, and Tableau for interactive visualization. The goal is to generate actionable insights and dashboards for stakeholders.
File structure
• Top-level artifact:
• data/financial_loan.csv → raw source data
• sql_queries/01_*.sql → ETL and aggregation scripts
• assets/*.png → finalized dashboard screenshots
• docs/… → problem statement and query documentation
• No application servers or traditional front/back ends—this is an ETL + visualization pipeline.
• No container, infra as code, or deployment manifests—this runs ad hoc or via a simple script/IDE.
Key insights from README
• Methodology: three stages—Data Extraction & Aggregation (SQL), EDA, and Visualization in Tableau.
• Users interact with a Tableau Public dashboard (assets store screenshots).
• Stakeholders use filters on loan status, purpose, grade.
a. Main components
1. Raw Data Layer
– CSV file (financial_loan.csv) stored on disk or data lake
2. ETL/SQL Processing Layer
– SQL Scripts (01_kpi_summary.sql, …) executed against the CSV (via a SQL engine like SQLite, Postgres, or a data-warehouse)
– Intermediate tables/views for cleaned and aggregated data
3. Documentation & Queries
– docs/ contains problem statement and detailed query explanations
4. Visualization Layer
– Tableau Desktop / Tableau Public
– Interactive dashboard hosted on Tableau Public
5. Assets
– Static images of dashboards for README or reporting
6. End Users / Stakeholders
– Business analysts, risk managers viewing dashboards
b. Relationships & interactions
c. Architectural patterns & design principles
– ETL pipeline: separate extraction (raw CSV), transformation (SQL), load into BI tool
– Modular SQL scripts for each analytical objective (KPI summary, status analysis, dimensional analysis)
– Single source of truth: the processed tables feed all dashboard visuals
– Documentation-driven: query docs and problem statement guide pipeline development
d. Technologies, frameworks, libraries
– Storage: flat file (CSV) or data-warehouse
– Processing: SQL engine (specify Postgres/SQLite/big query as used)
– BI: Tableau Desktop & Tableau Public
– Version control: Git (via .gitignore, LICENSE)