Automated Business Data Cleaning & Profit Dashboard

Upload a messy sales/expense spreadsheet and instantly get a clean profit dashboard plus a downloadable Excel workbook and PDF report.

Built as a portfolio MVP to show how a small business or freelancer can turn their bookkeeping export into actionable insights — no manual cleanup, no formulas, no logins.

Live Demo

Frontend: https://business-expense-sales-dashboard.vercel.app
Backend Health Check: https://business-expense-sales-dashboard.onrender.com/health

Features

• Drag-and-drop CSV / XLSX / XLS upload with size and type validation.
• One-click sample dataset so anyone can demo the app without uploading.
• Automatic cleaning — header normalization, currency parsing, date parsing, duplicate removal, missing-category fill-in, income/expense classification.
• Profit dashboard — total sales, total expenses, net profit, profit margin, and transaction count.
• Charts — monthly trend (sales / expenses / profit) and a category breakdown donut.
• Top products / services ranked by revenue contribution.
• Cleaned data preview in a responsive table.
• Data quality summary — duplicates removed, empty rows dropped, invalid amounts skipped, missing categories filled.
• Downloadable Excel & PDF reports generated server-side.

Tech stack

Layer	Tools
Frontend	React 19, Vite, Tailwind CSS v4, Recharts, Axios, lucide-react
Backend	FastAPI, Uvicorn, pandas, openpyxl, ReportLab
Reports	Excel (openpyxl) + PDF (reportlab) generated on each analysis

Screenshots

Screenshots of the dashboard live in screenshots/. Drop your own captures there (e.g. dashboard.png, upload.png) and embed them above the Features section once you have them.

Project structure

business-expense-sales-dashboard/
├── backend/                FastAPI app, analysis pipeline, generated reports
│   ├── app/
│   │   ├── main.py         API entry: /health, /api/analyze, /api/download/*
│   │   ├── services/       cleaner, analyzer, categorizer, exporters
│   │   └── utils/          column mapping for messy headers
│   ├── generated/          per-request Excel + PDF files (created at runtime)
│   ├── requirements.txt
│   └── runtime.txt         pinned Python version (used by Render)
├── frontend/               React + Vite dashboard
│   ├── src/
│   │   ├── api/client.js   axios + base URL + sample CSV + error formatter
│   │   ├── components/     FileUpload, SummaryCards, charts, tables, …
│   │   ├── pages/Dashboard.jsx
│   │   └── index.css       Tailwind v4 entry
│   ├── .env.example
│   └── package.json
├── sample-data/            example messy spreadsheets for testing
│   ├── messy-business-data.csv
│   ├── ecommerce-sample.csv
│   └── freelancer-income-expense.csv
└── screenshots/            put dashboard screenshots here for the README

Local setup

Prerequisites

• Python 3.11.x
• Node 18+ (Node 20 LTS recommended)
• Windows PowerShell, macOS Terminal, or Linux shell

1. Backend

cd backend
python -m venv venv
.\venv\Scripts\Activate.ps1     # macOS/Linux: source venv/bin/activate
pip install -r requirements.txt
uvicorn app.main:app --reload

The API is now at http://127.0.0.1:8000. Quick check:

curl http://127.0.0.1:8000/health
# → {"status":"ok"}

API endpoints:

Method	Path	Purpose
GET	/health	Liveness check
POST	/api/analyze	Multipart upload → analysis JSON
GET	/api/download/excel/{file_id}	Download generated Excel report
GET	/api/download/pdf/{file_id}	Download generated PDF report
GET	/favicon.ico	204 no-content

2. Frontend

cd frontend
copy .env.example .env.local        # macOS/Linux: cp .env.example .env.local
npm install
npm run dev

Open http://127.0.0.1:5173.

Environment variables

frontend/.env.local:

VITE_API_BASE_URL=http://127.0.0.1:8000

In production, set this to your deployed backend URL (e.g. https://your-api.onrender.com).

The backend reads an optional FRONTEND_ORIGIN env var to extend its CORS allow-list (the local Vite port is allowed by default).

Sample data

Three ready-made CSVs live in sample-data/:

File	Use case
messy-business-data.csv	Tiny demo with mixed currency formatting
ecommerce-sample.csv	E-commerce store: orders, ad spend, shipping
freelancer-income-expense.csv	Freelancer: client invoices, software, office

You have three ways to try the app without your own file:

1. Click "Try sample data" in the upload panel — the UI ships a small built-in demo CSV to /api/analyze.
2. Upload one of the files in sample-data/ from the upload panel.
3. POST a sample file directly to the API:

   curl.exe -F "file=@sample-data/ecommerce-sample.csv" http://127.0.0.1:8000/api/analyze

Test checklist

Run through this list before shipping a build or recording a portfolio demo.

Backend smoke tests (with the API running on port 8000):

• GET /health returns {"status":"ok"}.
• POST /api/analyze with sample-data/messy-business-data.csv returns a JSON body that includes summary, monthly_trend, category_breakdown, top_products, preview_rows, data_quality, download_excel_url, and download_pdf_url.
• GET /api/download/excel/{file_id} downloads a valid .xlsx file.
• GET /api/download/pdf/{file_id} downloads a valid .pdf file.
• GET /api/download/excel/does-not-exist returns a clean 404.
• Uploading a .txt file returns a 400 with a friendly message.
• Uploading an empty file returns a 400 with a friendly message.

Frontend smoke tests (with both servers running):

• Hero, "How it works", and the upload card all render on first load.
• Clicking Try sample data populates the dashboard end-to-end: summary cards, monthly trend chart, category donut, top products, data quality card, and cleaned data preview.
• Uploading sample-data/ecommerce-sample.csv shows duplicate rows removed in the data quality card.
• Uploading sample-data/freelancer-income-expense.csv shows missing categories filled in the data quality card.
• Download Cleaned Excel and Download PDF Report are disabled before any analysis runs and become real download links after.
• Stopping the backend and clicking Try sample data shows a friendly "Cannot reach the backend" error card; restarting the backend lets a retry succeed.
• npm run build finishes without errors.

Deployment plan

Backend → Render

1. Push this repo to GitHub.
2. On Render, create a new Web Service from the repo.
3. Settings:
   • Root directory: backend
   • Build command: pip install -r requirements.txt
   • Start command: uvicorn app.main:app --host 0.0.0.0 --port $PORT
   • Environment: Python (Render reads runtime.txt → python-3.11.9)
   • Env var: FRONTEND_ORIGIN=https://your-frontend.vercel.app

Frontend → Vercel

1. On Vercel, import the same GitHub repo.
2. Settings:
   • Root directory: frontend
   • Build command: npm run build
   • Output directory: dist
   • Env var: VITE_API_BASE_URL=https://your-backend.onrender.com
3. Redeploy after the backend is live so the env var is baked into the build.

Portfolio business use case

Small businesses and freelancers usually keep their books in a spreadsheet that looks "good enough" but is hard to read at a glance:

• Currency mixed with text ("$1,200", "৳500", "(1000)")
• Different column names every month (Order Date vs Date vs Sale Date)
• Missing categories
• Duplicate or empty rows

This app is the "instant clean + profit report" tool that converts that messy export into a presentation-ready dashboard and a shareable PDF — useful for monthly reviews, year-end summaries, or quick client reports.

Future improvements

Kept deliberately out of the MVP — only worth adding when a real user needs them:

• Multi-currency support — detect currency per row and report totals per currency.
• Persistent history — a tiny SQLite store so a user can compare months or re-open a previous analysis without re-uploading.
• CSV download of the cleaned dataset (in addition to Excel + PDF).
• Date range filter on the dashboard for users with large files.
• AI-assisted categorization as an opt-in upgrade for messy descriptions the keyword categorizer can't classify.
• Auth + per-user history — only if there's a real customer asking.

License

MIT — use it freely for portfolio, learning, or as a starting point.