Prismfly

Decompose your spending, track your real purchasing power.

After 10 years of manually recording every transaction and reconciling all accounts every morning in Firefly III, the dataset is rich enough for deeper questions: how has my real cost of living evolved? Is my discretionary spending growing faster than my income? Can I sustain my investment plan while keeping a comfortable margin?

Prismfly crosses that granular spending data with income tax records to answer those questions — with an interactive dashboard and a markdown report you can feed to an LLM for further analysis.

Features

• Personal Inflation Index (ICVP): tracks how your cost of living evolves relative to a base year, compared against official CPI
• Lifestyle inflation elasticity: measures whether discretionary spending grows faster or slower than income
• Account-based savings tracking: net flow in main accounts as the most reliable savings metric
• Investment sustainability waterfall: monthly cascade from salary through needs, investment, and discretionary spending to free margin
• Interactive Plotly.js dashboard: 11 charts with zoom, pan, hover, legend toggle, range sliders, and stacked/grouped modes
• Markdown report: raw analytical data exportable for LLM-assisted analysis
• Shared expenses support: optional partner reimbursement tracking via Settle Up or similar
• Capital expenditure filtering: automatic exclusion via Firefly tags or description keywords

Requirements

• uv
• Firefly III instance accessible on your network
• .env file with FIREFLY_URL and FIREFLY_TOKEN
• personal.json with your tax data and account structure (see personal.example.json)

Usage

# Fetch latest data from Firefly III and generate dashboard + report
uv run main.py

# Regenerate outputs from cached data only (no network)
uv run main.py --no-fetch

# Force re-download of all years (not just current)
uv run main.py --force

Outputs:

• dashboard.html — interactive dashboard, open in any browser
• dashboard.md — raw metrics in markdown, for LLM analysis or external tools

For LLM-assisted discussion, provide both dashboard.md and CONTEXT.md (your manually maintained narrative context).

Project Structure

├── main.py               # CLI entry point
├── src/
│   ├── config.py         # Loads personal.json + public data (CPI)
│   ├── firefly.py        # Firefly III API client (read-only)
│   ├── extract.py        # Data extraction, pagination, and caching
│   ├── analyze.py        # Basket classification, ICVP, elasticity, waterfall
│   ├── dashboard.py      # Plotly.js HTML generation
│   └── report.py         # Markdown report generation
├── personal.json         # Your tax data, accounts, filters (gitignored)
├── personal.example.json # Template for personal.json
├── CONTEXT.md           # Narrative context for LLM analysis (manual, gitignored)
├── data/                 # Cached JSON (auto-generated, gitignored)
├── .env                  # API URL and token (gitignored)
├── dashboard.html        # Generated dashboard (gitignored)
└── dashboard.md          # Generated report (gitignored)

Configuration

All personal data lives in personal.json (gitignored). See personal.example.json for the full schema. Key sections:

Section	Purpose
tax_data	Annual gross/net salary by year
salary_sources	Employer names for income classification
main_accounts	Firefly account names + IDs for savings tracking
partner_*	Shared expense reimbursement config (optional)
basket_map	Firefly category → analytical basket mapping
capital_tags / capital_keywords	Capital expenditure detection
renovation_tags	One-off renovation exclusion from sustainability waterfall
shield_accounts / investment_accounts / other_accounts	Financial snapshot for the report

Dashboard

11 interactive charts (Plotly.js):

1. Quarterly spending — needs / discretionary / other, with stacked/grouped toggle
2. Monthly explorer by basket — range slider, stacked/grouped/100% modes
3. ICVP vs CPI — personal inflation vs official index
4. Elasticity scatter — income variation vs discretionary spending variation
5. Heatmap — average monthly spend by basket and year
6. Distribution (%) — spending composition by basket
7. Dining & leisure — EUR/month and % of salary (lifestyle inflation indicator)
8. Monthly savings rate — with 3-month moving average
9. Main accounts delta — annual net flow
10. Income composition — salary, reimbursements, inheritance, other
11. Sustainability waterfall — full per-basket breakdown from salary to free margin

Caching

Transaction data is cached as JSON in data/. Closed years are never re-fetched unless --force is used. The current year is always refreshed. Account balances are refreshed on every fetch run.

Maintenance

Tax data

Update personal.json with each new tax year. Update IPC_ACUM / IPC_INTER in config.py with official CPI data.

Categories and filters

• Subscription reclassification: subs_education / subs_technology in personal.json
• Renovation one-offs: tag transactions in Firefly with the tags listed in renovation_tags
• Capital expenditures: tag with capital_tags or add description keywords to capital_keywords

Narrative context

Update CONTEXT.md when personal circumstances, investment strategy, or major financial events change.

License

MIT