Standardizing non-standard cyclical data (e.g., Luni-Solar / Panchang metrics) as features in execution & signal pipelines

External question — Quantitative Finance Stack Exchange Author: Krishna Moorthy M Original post: https://quant.stackexchange.com/questions/85841 License: CC BY-SA 4.0 — https://creativecommons.org/licenses/by-sa/4.0/ Adaptation: HTML converted to plain text; contact email addresses removed. I am designing a quantitative trading strategy inside the XtraAlgoQ PULSE architecture. The strategy relies on non-standard alternative signals based on luni-solar calendar metrics (traditionally categorized as Panchang variables—such as solar/lunar longitude differences, tithi durations, and planetary sidereal angles) to evaluate potential intraday regime changes and volume/volatility anomalies. While the XtraAlgoQ PULSE platform handles high-frequency data and standard market parameters via direct feeds, incorporating continuous real-time lunar/solar angles creates specific quantitative modeling challenges: Stationarity & Feature Normalization: Panchang metrics rely on cyclic/angular coordinates (0°–360°) and non-uniform time intervals (e.g., varying tithi durations rather than fixed-width time bars). What are the standard methods for transform-encoding these periodic variables into stationary features suitable for real-time statistical signals (e.g., Sine/Cosine positional encodings vs. Fourier expansion)? Backtesting & Look-Ahead Bias: Because calendar-based metrics depend on geographic observer location and astronomical ephemeris algorithms, how can one ensure zero look-ahead bias when generating historical tick-level signals for backtesting? Statistical Significance Testing: Given the high degree of noise in daily financial time series, what quantitative framework (e.g., Permutation tests, Deflated Sharpe Ratio, or Bootstrapping) is best suited to verify whether any observed correlation between these cyclic metrics and intraday volatility is statistically significant rather than an artifact of multiple testing? How do quantitative researchers effectively map continuous non-standard temporal variables into quantitative pipeline architectures without introducing severe overfitting? https://www.youtube.com/watch?v=JVtUcM1sWQw This video provides an introductory breakdown of how quantitative finance uses probability distributions, statistical modeling, and alternative data features to build systematic trading strategies, offering helpful context for structuring quantitative signals. My App: https://xtraalgoq-pulse.ai.studio