AI Data Analyst Platform
Conversational analytics platform designed as a "ChatGPT for data analysis": it turns natural-language questions into safe SQL, automated analysis, interactive visualizations, insights, and predictive outputs.
Stack
Objective
Build an end-to-end analytics product where users can upload data, ask business questions in natural language, and automatically receive SQL, analysis, charts, recommendations, and predictive outputs.
Architecture
- Streamlit frontend for dataset upload, natural-language questions, charts, and report downloads.
- FastAPI backend exposing query, dataset registration, insight history, and chart endpoints.
- Specialized agent system for orchestration, SQL generation, analysis, visualization, and insights.
- Data layer compatible with CSV, SQLite, and PostgreSQL.
- Machine learning services for forecasting, clustering, and anomaly detection.
Key capabilities
Natural Language to SQL
Interprets business questions and generates safe read-only SQL queries.
Automated EDA
Builds profiling, descriptive statistics, missing-value reports, correlations, and outlier checks automatically.
Smart visualization
Selects chart types automatically and delivers interactive Plotly outputs.
Prediction and anomalies
Extends analysis with forecasting, clustering, and anomaly detection workflows.
Screenshots
Outcome
This project demonstrates product thinking, backend architecture, LLM integration, visualization, operational security, and VPS deployment. It is a strong flagship piece because it shows the ability to build complete analytics products, not just isolated dashboards.