README.md

metadata

title: My Streamlit App
emoji: 🚀
colorFrom: blue
colorTo: green
sdk: streamlit
sdk_version: 1.39.0
app_file: app.py
pinned: false

AI Spending Analyser (Streamlit)

Features

• Synthetic dataset (~900 rows) across ~1 year with realistic variability
• Filters: date range, categories, merchant query
• Metrics: total, average monthly, max/min transaction
• Charts: daily trend (line), spend by category (bar), payment methods (donut)
• AI summary: OpenAI GPT if OPENAI_API_KEY exists, else deterministic heuristic summary
• CSV download of filtered data

Quickstart

1. Install python -m venv .venv . .venv/Scripts/activate # Windows PowerShell pip install -r ai_spending_analyser/requirements.txt

2. Run locally streamlit run ai_spending_analyser/app.py

3. (Optional) Enable OpenAI summaries

PowerShell

$env:OPENAI_API_KEY = "sk-..."

Deploy to Streamlit Cloud

1. Push this folder to a GitHub repo.
2. On Streamlit Cloud, create a new app pointing to ai_spending_analyser/app.py.
3. Add OPENAI_API_KEY as a secret if you want AI summaries.

Libraries

• streamlit
• pandas
• numpy
• plotly
• openai (optional)
• ollama (optional; for free local LLM)

Local LLM (Ollama)

1. Install Ollama: https://ollama.com
2. Run Ollama and pull a small model, e.g.: ollama pull llama3.2
3. In the app sidebar, set Engine to "Ollama" and (optionally) model to "llama3.2".
4. No API keys needed; runs fully offline.

Notes

• The app gracefully handles empty filters by showing an info message.
• Regenerate button synthesizes a fresh dataset.