Spaces:
Sleeping
Sleeping
| title: BizIntel AI Ultra | |
| emoji: 📊 | |
| colorFrom: indigo | |
| colorTo: blue | |
| sdk: docker | |
| app_file: app.py | |
| pinned: false | |
| # 📊 BizIntel AI Ultra – Advanced Analytics & Agentic BI | |
| **BizIntel AI Ultra** turns any data table—whether a quick CSV upload or a live | |
| database table—into an interactive **insight & forecasting workspace** powered | |
| by: | |
| * **Google Gemini 1.5 Pro** ⇒ natural-language strategy briefs | |
| * **ARIMA time-series pipeline** ⇒ 90-step (or 3-period) forecasts with | |
| model explainability, back-testing, residual ACF & CI bands | |
| * **Plotly Dark™** trend, forecast & EDA visuals | |
| * **Streamlit** front-end running inside a Hugging Face Space | |
| * **Modular tools** (`csv_parser`, `plot_generator`, `forecaster`, | |
| `visuals`, `db_connector`) for quick reuse in your own projects | |
| --- | |
| ## 🚀 Live Demo (Hugging Face Space) | |
| <https://huggingface.co/spaces/your-handle/BizIntel_AI> | |
| Upload a CSV or Excel (≤ 500 MB) **or** connect to PostgreSQL / MySQL / | |
| SQL Server / Oracle via a SQLAlchemy URI, select a date column and a numeric | |
| metric, and watch: | |
| 1. Trend chart | |
| 2. 90-step forecast with confidence fan, summary stats & interpret-ability | |
| 3. Gemini-generated strategy report | |
| 4. Optional histograms, scatter matrices, correlation heat-maps | |
| --- | |
| ## ✨ Features | |
| | Area | Details | | |
| | ---- | ------- | | |
| | **Data ingest** | CSV, XLS/XLSX, or DB fetch (`sqlalchemy` URI) | | |
| | **Sampling** | Auto-samples giant CSVs (≥ 1 M rows) to avoid OOM | | |
| | **Quick scan** | Markdown dataset overview: schema, missing %, numeric `describe()` | | |
| | **Trend** | Interactive line + markers (Plotly Dark) | | |
| | **Forecast** | ARIMA(1,1,1) auto-freq, 90-step (daily) or 3-step (monthly, etc) | | |
| | | CI fan, residual ACF plot, 20 % back-test (MAPE, RMSE) | | |
| | | Seasonal decomposition (toggle) | | |
| | **Explainability** | Model summary table, plain-English coef notes | | |
| | **Gemini 1.5 Pro** | 5 insights, 3 strategies, risks, extra-visuals → Markdown + download | | |
| | **EDA** | Histograms (with KDE), box / violin plots, scatter-matrix, corr heat-map | | |
| | **Static images** | All figures also saved to `/tmp/*.png` for downstream reports | | |
| | **Extensible** | Each visual is a standalone helper in `tools/` | | |
| --- | |
| ## 🛠 Local Dev Setup | |
| ```bash | |
| # 1. Clone | |
| git clone https://github.com/your-handle/BizIntel_AI.git | |
| cd BizIntel_AI | |
| # 2. Python env | |
| python -m venv .venv | |
| source .venv/bin/activate | |
| pip install -r requirements.txt | |
| # 3. Gemini key | |
| export GEMINI_APIKEY="your_google_ai_studio_key" | |
| # 4. Run | |
| streamlit run app.py | |
| Note : ARIMA & seasonal decomposition come from | |
| statsmodels; Plotly PNG export needs kaleido. | |
| These are included in requirements.txt. | |
| 🔌 Database Connect String Examples | |
| text | |
| Copy code | |
| # PostgreSQL | |
| postgresql+psycopg2://user:password@host:port/dbname | |
| # MySQL | |
| mysql+pymysql://user:password@host/dbname | |
| # SQL Server | |
| mssql+pyodbc://user:password@host:1433/db?driver=ODBC+Driver+17+for+SQL+Server | |
| # Oracle | |
| oracle+cx_oracle://user:password@host:1521/?service_name=ORCL | |
| 🌐 Deploying to Hugging Face Spaces | |
| Space ▸ Create ▸ Docker template | |
| Push repo (Dockerfile, app.py, tools/, requirements.txt) | |
| Add a secret → GEMINI_APIKEY | |
| Space builds ; enjoy! | |
| 🗺 Project Structure | |
| Copy code | |
| ├─ app.py | |
| ├─ requirements.txt | |
| ├─ Dockerfile | |
| ├─ tools/ | |
| │ ├─ csv_parser.py | |
| │ ├─ plot_generator.py | |
| │ ├─ forecaster.py | |
| │ ├─ visuals.py | |
| │ └─ db_connector.py | |
| └─ README.md | |
| 🤝 Contributing | |
| Issues & PRs are welcome! | |
| Please run black -l 88 and ruff before submitting. | |
| 📜 License | |
| MIT © 2025 BizIntel AI Ultra Team |