Description
Decomposes a behavioural time series into frequency components to reveal which tempos dominate activity, from daily habits to rapid micro-fluctuations.
Fourier analysis represents a signal as a sum of sinusoids; the power spectrum shows how much energy sits at each frequency. Vivek Gopinath's Behavioral Spectral Framework treats it as an 'X-ray of habit': dominant low frequencies indicate stable long-term habits, strong high-frequency energy indicates micro-fluctuation, distraction or reactivity, and multiple harmonics suggest layered goals. Because human behaviour is non-stationary, the playbook recommends windowed transforms (STFT) or treating frequency change itself as a dependent variable, recurrence strength over strict periodicity, and broad-band power ratios rather than single peaks. In the fintech case, STFT showed a dominant band of 0.15–0.25 Hz and high-frequency bursts of 0.8–1.2 Hz at KYC and form-filling. Interpretability is a strength: frequency maps to pace, amplitude to intensity and phase to timing, but amplitude is not the same as engagement and must be read with context.
Attributes
- Discipline
- Information Theory, Physics & Mathematics
- Loss type
- Unmigrated
- Stage
- Post-launch
- Data
- Quantitative
- Effort
- Days
Steps
- Frame the signal (e.g. interaction density per minute) with time resolution matched to the hypothesised rhythm.
- Resample or interpolate small gaps, keeping long gaps as meaningful silence, or switch to Lomb-Scargle for irregular events.
- Compute the FFT for global periodicity, or STFT windows to see change over time.
- Extract dominant frequencies, broad-band power ratios (slow versus fast) and relative energy drift ΔP/P.
- Compare spectra across cohorts (novice versus expert, calm versus stressed).
- Interpret peaks with context metadata and qualitative evidence before reporting.
Sources
- Vivek Gopinath (2025), From Stars to Signals: Translating Matteo Paz's Fourier-Wavelet Model into Behavioral and UX Research Methodology
- Matteo Paz (2024), A Sub-Millisecond Fourier and Wavelet Based Model to Extract Variable Candidates from the NEOWISE Single-Exposure Database
Also known as
- FFT of behaviour
- Short-Time Fourier Transform (STFT)
- Behavioural X-ray
Categories
- Information Theory, Physics & Mathematics