jquantstats¶

Portfolio analytics for quants — built on Polars, powered by Plotly.

Quick Links: 📚 Repository • 📦 PyPI • 🐛 Issues • 💬 Discussions

📋 Overview¶

jQuantStats is a Python library for portfolio analytics that helps quants and portfolio managers understand their strategy performance in depth. It provides two complementary entry points: a Portfolio route that works directly from price and position data, and a Data route for arbitrary return streams. All analytics, visualisations, and HTML reports are available from either entry point.

🚀 Installation¶

pip install jquantstats

Python 3.11+ is required. Add optional extras as needed:

pip install jquantstats[plot]   # static chart export (kaleido)
pip install jquantstats[web]    # FastAPI web server

💻 Quick Start¶

import polars as pl
from jquantstats import Portfolio

pf = Portfolio.from_cash_position(
    prices=prices_df,      # pl.DataFrame: date + asset columns
    cash_position=pos_df,  # pl.DataFrame: date + asset columns (£ amounts)
    aum=1_000_000,
)

pf.stats.sharpe()          # {"AAPL": 1.34, "MSFT": 0.91, "portfolio": 1.21}
pf.plots.snapshot()        # interactive NAV + drawdown dashboard
pf.report.to_html()        # full HTML tearsheet

import polars as pl
from jquantstats import Data

data = Data.from_returns(
    returns=returns_df,     # pl.DataFrame: date + return columns
    benchmark="Benchmark",  # optional column name
    rf=0.0,
)

data.stats.sharpe()
data.stats.sortino()
data.stats.annual_breakdown()
data.plots.monthly_heatmap()
data.reports.to_html()

🔬 Portfolio Route — Why It Matters¶

Working from positions rather than returns unlocks analysis that is impossible from a return series alone:

• pf.lag(n) — execution-delay analysis: simulate T+1, T+2 fills and see the Sharpe impact
• pf.tilt / pf.timing — decompose performance into allocation skill vs. timing skill
• pf.turnover / pf.turnover_summary() — daily and weekly turnover analytics
• pf.trading_cost_impact(max_bps=20) — sweep transaction-cost sensitivity in one call

pf_t0 = pf           # ideal T+0 execution
pf_t1 = pf.lag(1)    # signal fires today, fills tomorrow

print(pf_t0.stats.sharpe())  # {"portfolio": 1.34}
print(pf_t1.stats.sharpe())  # {"portfolio": 1.28}

fig = pf.plots.lead_lag_ir_plot(start=-5, end=10)

See Getting Started for a complete walkthrough.

📊 jQuantStats vs QuantStats¶

See the Migration Guide for a complete API mapping.

📄 Next Steps¶