How tripathi-22/cryptocurrency-insights-dashboard works
Project type and purpose
• This is a data‐analytics & BI pipeline rather than a traditional web application. It ingests raw crypto CSVs, performs ETL and machine learning in Python, loads results into a MySQL database, and finally builds an interactive Power BI dashboard.
File structure
• Top‐level files:
• README.md — project overview and instructions
• final.pbix — Power BI dashboard file
• merged_data.csv — consolidated dataset output from ETL
• python.py — main ETL + ML + database integration script
• python_visualization (directory) — likely holds Python‐generated plots or notebooks
• No frontend/backend folders per se; instead you have an ETL/ML component and a BI visualization component.
Extract key insights from the README
• Four sequential stages:
1. Data Preparation (merge & clean 20+ coin CSVs)
2. Machine Learning (KNN & Decision Tree)
3. MySQL Integration (schema design, data import, SQL views)
4. Power BI Dashboard (interactive visuals, DAX, slicers)
• Tech stack: Python (pandas, seaborn, matplotlib, scikit-learn), MySQL, Power BI.
a. Main components
• Data Sources: original per‐coin CSV files
• ETL & ML Engine: python.py & python_visualization folder
• Storage Layer: MySQL database
• BI Layer: Power BI desktop (final.pbix)
• Users/Consumers: Analysts interacting with the dashboard
b. Relationships & interactions
1. CSV files feed into the ETL script.
2. ETL cleans/merges data and outputs merged_data.csv.
3. Python script trains ML models on merged_data.csv, produces prediction output.
4. Python script connects to MySQL, creates tables & views, and writes both raw and ML results.
5. Power BI connects to MySQL views to build visuals and filters.
6. End users consume the Power BI report.
c. Architectural patterns & principles
• Layered pipeline:
• Ingestion layer (file I/O)
• Processing layer (data cleaning + feature engineering)
• Analytics layer (machine learning)
• Persistence layer (database)
• Presentation layer (dashboard)
• Modular scripts: single python.py orchestrates multiple sub‐tasks
• Separation of concerns: BI layer purely reads from DB, no direct file processing
d. Key technologies/frameworks
• Python (pandas, seaborn, matplotlib) for ETL & visual exploration
• scikit-learn for ML
• mysql-connector-python for DB connectivity
• MySQL engine for storage, indexing, and SQL views
• Power BI for front‐end visualization, DAX, slicers, and themes