SmartDash - AI Powered Dashboard

👉 An AI-ready, user-friendly analytics dashboard for small businesses.

SmartDash is a lightweight web app that turns raw sales CSV files into instant insights.

Use cases:

• Small shops analyzing monthly sales
• Quick product/region performance checks

✨ Key Features

• Auto Column Detection for Date, Product, Quantity, Price/Revenue (with manual selectors if needed)
• Interactive Filters: Category, Region, Sales Rep (when available in data)
• KPI Cards: Total Revenue, Units Sold, Avg Order Value, Top Product/Region
• Charts: Sales trend (monthly), Top 5 Products, Revenue by Region
• Insights: Quick bullets(seasonality, best weekday, best rep, low performers)
• Export: Filtered CSV + styled PDF report (light-blue theme + logo placeholder)
• Mobile-friendly layout with scrollable preview
• Session-aware: Remembers user selections across uploads in the same session

🖼 Demo Screens

Homepage

Loaded Data

🎯 Quick Start: Upload your sales CSV, explore insights, and download reports—all in your browser!

🧾 Dataset Format & Column Detection

SmartDash expects a CSV with typical sales columns. It automatically tries to detect these by name:

• Date: Sale_Date, Date, Order_Date, Transaction_Date
• Product: Product_Name, Product, Product_ID, Item
• Category (optional): Category, Product_Category
• Region (optional): Region, State, Location
• Sales Amount: Sales_Amount, Total, Amount, Sales
• Quantity: Quantity, Qty, Units, No. of items
• Unit Price: Unit_Price, Price, Item Price
• Unit Cost (optional): Unit_Cost, Cost
• Sales Rep (optional): Sales_Rep, Salesperson

If Sales Amount is missing, SmartDash computes Revenue = Quantity × Unit Price (if both exist).
If auto-detection fails, the always-visible selectors let you pick the right columns.

Tip: Use a sales_sample.csv for easy demos - Download

🧭 How to Use the Dashboard

1. Upload CSV — drag & drop your file.
2. (Optional) Adjust column selectors — if the app couldn’t auto-detect correctly.
3. Apply filters — pick Category/Region/Sales Rep to narrow results.
4. View KPIs — instant totals and top performers.
5. Explore charts — sales trend, top products, revenue by region.
6. Read Insights — quick bullets; detailed findings.
7. Export — download filtered CSV or a styled PDF report.

🧠 Insights Explained

SmartDash shows Quick Insights by default

• Quick Insights (bulleted):
  • Top product & top region
  • Average order value
  • Month-over-month change
  • Profit & margin (when cost data exists)

🧾 Exporting Data & Reports

• CSV Export: Downloads exactly what you’ve filtered in the app.
• PDF Export: Branded, light-blue theme with sections:
  • Header with logo + app title

The PDF uses ReportLab.

To use Offline:

📦 Requirements

List of key Python packages (also in requirements.txt):

streamlit
pandas
numpy
reportlab
altair
datetime
# If you read Excel files too:
# openpyxl
# If you render matplotlib charts into the PDF:
# matplotlib

> **Python version:** 3.9+ recommended. Works on Windows/macOS/Linux.

---

## 🧑‍💻 Getting Started (Local)

1) **Clone the repo**
```bash
git clone https://github.com/yourusername/smartdash.git
cd smartdash

2. Create a virtual environment (recommended)

# Windows
python -m venv .venv
.venv\Scripts\activate

# macOS/Linux
python3 -m venv .venv
source .venv/bin/activate

3. Install dependencies

pip install -r requirements.txt

4. Run SmartDash

streamlit run app.py

5. Open the app
   Streamlit will show a local URL like http://localhost:8501. Open it in your browser.

🧰 Troubleshooting

• “streamlit: command not found”
  Activate your virtual env or reinstall: pip install streamlit.

• “File does not exist: app.py”
  Make sure you’re in the folder that contains app.py when running streamlit run app.py.

• “TypeError: unsupported operand type(s) for /: ‘str’ and ‘int’”
  Ensure numeric columns are converted with pd.to_numeric(..., errors="coerce") and handle NaN with .fillna(0).

• Auto-detection wrong?
  Use the column selectors to manually assign columns.

• Date not parsing?
  The app tries multiple formats and uses errors="coerce". You can pre-clean dates in your CSV if needed.

📄 License

This project uses the MIT License — open and permissive.
See the LICENSE file for full text.