Trading Audit Dashboard

Forensic trading performance audit — from any broker statement to an institutional-grade HTML dashboard in seconds.

Supports Trading 212, Interactive Brokers, eToro, MetaTrader 4/5, Binance Futures, NSE/NFO (Indian markets), and any generic CSV export.

---

---

What It Does

Upload any broker trading statement (PDF, CSV, XLSX, or JSON) and get a self-contained dark-theme HTML dashboard with:

• Executive summary — Net P&L, win rate, profit factor, drawdown, Sharpe ratio
• Equity curve with monthly P&L breakdown
• Strategy reverse-engineering — trade archetype, edge analysis, fat-tail risk
• Behavioral forensics — session analysis, revenge trading detection, tilt patterns
• Advanced analytics — Kelly sizing, risk of ruin, rolling windows, hourly P&L heatmap
• Monte Carlo projections — 10,000 simulations, bull/base/bear scenarios
• Quantified recommendations — up to 5 actionable rule changes with expected impact

---

Features

7-Phase Analysis

Phase	Description
1. Executive Dashboard	Vital stats: net P&L, win rate, profit factor, max drawdown, expectancy, Sharpe
2. Strategy Reverse-Engineering	Archetype classification, edge identification, fat-tail risk
3. Behavioral Forensic Audit	Session/DOW analysis, revenge factor, holding patterns
4. Advanced Analytics (v2.0)	Kelly sizing, risk of ruin, rolling performance, hourly heatmap, streak analytics
5. Statistical Forecasting	Monte Carlo simulation (10,000 runs), 3-scenario projections
6. Professional Recommendations	Up to 5 quantified, actionable improvements
7. JSON Export	Complete stats object (70+ fields) for downstream use

v2.0 New Metrics

• Wilson Confidence Interval on win rate (95% CI)
• Kelly Criterion — Full Kelly + Half Kelly position sizing
• Risk of Ruin — at current sizing and 2× leverage
• Drawdown Recovery — estimated trades and days to recover
• Loss Streaks & Tilt Detection — consecutive loss patterns
• Rolling Period Comparison — 30/60/90-day performance windows
• Hourly P&L Heatmap — UTC 0–23 heat map of best/worst trading hours
• Multi-Instrument Matrix — per-symbol Sharpe, win rate, and P&L
• Transaction Cost / Swap Drag — quantified cost as % of gross profit
• Streak & Momentum Analytics — hot/cold windows, post-win overconfidence

---

Supported Brokers

Broker	Format	Notes
Trading 212	PDF	CFD statement, auto-detects closing entries
Interactive Brokers (IBKR)	CSV Flex Query	Full position and P&L columns
eToro	CSV / XLSX	Realised P&L export
MetaTrader 4 / 5	CSV / HTML	Detailed statement or history export
Binance Futures	CSV	Closed position history
NSE / NFO (India)	CSV	F&O, futures, and options trades
Generic CSV	CSV	Column-name heuristics for any broker

---

Quick Start — Run with Demo Data

Prerequisites

pip install pdfplumber openpyxl pandas

Step 1: Clone the repository

git clone https://github.com/vishalmdi/trading-audit-dashboard.git
cd trading-audit-dashboard

Step 2: Parse the demo trades

The demo dataset (demo/intraday_trading_log.csv) contains 4,155 real-style Indian F&O trades across NSE futures and options (Jan 2023 – Dec 2024).

python scripts/parse_trades.py demo/intraday_trading_log.csv
# Output: parsed_trades.json

Step 3: Compute statistics

python scripts/compute_stats.py parsed_trades.json
# Output: trading_stats.json  (70+ metrics)

Step 4: View the pre-built demo dashboard

Open the included demo output directly in your browser:

# macOS
open demo/trading_audit_dashboard.html

# Linux
xdg-open demo/trading_audit_dashboard.html

# Windows
start demo/trading_audit_dashboard.html

The dashboard shows the full analysis of the demo dataset — equity curve, Monte Carlo projections, behavioral forensics, Kelly sizing, and all 7 phases.

---

Using as a Claude Skill

This project ships as a Claude skill (.skill file) that can be installed directly into Claude.

Install

1. Download trading-audit-dashboard.skill
2. Open Claude → Settings → Skills → Import Skill
3. Import the .skill file

Trigger Phrases

Once installed, Claude activates this skill automatically when you say:

• "Analyze my trades"
• "Audit my trading history"
• "Create a trading dashboard"
• "Review my performance"
• "P&L analysis"
• "Trading report"
• Or simply upload a broker statement file

Claude will parse the file, compute all statistics, and produce a self-contained HTML dashboard.

---

File Structure

trading-audit-dashboard/
├── README.md                          This file
├── LICENSE                            MIT
├── SKILL.md                           Claude skill definition (v2.0)
├── trading-audit-dashboard.skill     Packaged skill archive (ZIP)
│
├── assets/
│   └── dashboard_template.html       Dark-theme HTML/CSS/JS dashboard template
│                                      (Chart.js, 44+ template placeholders)
│
├── references/
│   ├── broker_formats.md             Broker parsing rules and column mappings
│   └── statistics.md                 Statistical formulas and vulnerability thresholds
│
├── scripts/
│   ├── parse_trades.py               Universal broker statement parser
│   │                                  Supports PDF/CSV/XLSX, auto-detects broker
│   └── compute_stats.py              Statistics engine — 70+ metrics, v2.0 analytics
│
└── demo/
    ├── intraday_trading_log.csv      Sample dataset: 4,155 Indian F&O trades
    ├── trading_audit_dashboard.html  Pre-built demo dashboard output
    └── screenshot.png                Dashboard preview

---

Tech Stack

Component	Technology
PDF parsing	pdfplumber
XLSX parsing	openpyxl, pandas
Statistics	Pure Python (numpy-free for portability)
Dashboard UI	HTML5 + CSS3 + JavaScript
Charts	Chart.js 4.4
Fonts	Space Mono, Bebas Neue, DM Serif Display (Google Fonts)

---

Contributing

Pull requests are welcome. For major changes, please open an issue first to discuss what you'd like to change.

Areas for contribution:

• New broker parsers (add to scripts/parse_trades.py)
• Additional statistical metrics (add to scripts/compute_stats.py + references/statistics.md)
• Dashboard UI improvements (edit assets/dashboard_template.html)

---

License

MIT © 2026 vishalmdi