screen-factors

Factor Screening

You build candidate lists from data. Screening is not a backtest.

Confirm Inputs

Ask for:

• market and symbols
• factor definitions
• rebalance date or date range
• thresholds or ranking rules
• required data provider
• how to handle missing values

Inspect Available Indicators

uv run python - <<'PY'
import oxq
print(sorted(oxq.list_indicators()))
PY

Financial indicator classes include names such as PE, PB, BP, EP, ROEChange, NetProfitMargin, AccrualRatio, CashFlowRatio, MarketCap, TurnoverRate, and PowerRatio. Verify required input columns before computing them.

Price-Based Screening Pattern

import pandas as pd

from oxq.data.market import LocalMarketDataProvider
from oxq.indicators import NdayReturn, RollingVolatility

symbols = ["AAPL", "MSFT", "GOOGL"]
market = LocalMarketDataProvider(data_dir="/path/to/parquet")

rows = []
for sym in symbols:
    bars = market.get_bars(sym, "2020-01-01", "2024-12-31")
    momentum = NdayReturn().compute(bars, column="close", period=60).iloc[-1]
    volatility = RollingVolatility().compute(bars, column="close", period=20).iloc[-1]
    rows.append({"symbol": sym, "momentum_60": momentum, "vol_20": volatility})

screen = pd.DataFrame(rows).dropna()
screen["score"] = screen["momentum_60"].rank(pct=True) - screen["vol_20"].rank(pct=True)
candidates = screen.sort_values("score", ascending=False).head(10)

Financial Screening

For financial fields, use the factor data layer and inspect returned columns. Do not assume all financial indicators can compute from OHLCV bars alone.

Red Lines

• Do not call a screened list a validated strategy.
• Do not ignore missing factor values.
• Do not mix A-share and US data providers in one score without explaining it.
• Do not use future financial statement publication dates.