RateOfChangeEvents

Intervals where a signal changes faster than a rate limit, rate statistics per window, and sudden step jumps.

ts_shape.events.engineering.rate_of_change · Engineering guide · API reference

Example

import ts_shape
from ts_shape.events.engineering.rate_of_change import RateOfChangeEvents

df = ts_shape.make_timeseries(["flow_rate"], n_points=3600, freq="1s", baseline=50.0, noise=0.1)
df.loc[1800:, "value_double"] += 20.0  # step

roc = RateOfChangeEvents(df, "flow_rate")
rapid = roc.detect_rapid_change(threshold=0.5)
stats = roc.rate_statistics(window="15min")
result = roc.detect_step_changes(min_delta=10.0, max_duration="5s")
print(result[["systime", "value_before", "value_after", "delta"]])

Methods

Method	Returns
detect_rapid_change(threshold)	intervals with |rate| ≥ threshold: start, end, max_rate, direction
rate_statistics(window='1h')	per window: mean_rate, std_rate, max_rate, min_rate
detect_step_changes(min_delta, max_duration='5s')	jumps of at least min_delta within max_duration: time (after the jump), value_before, value_after, delta, duration

Notes

• Rates are in signal units per second, computed between consecutive samples. Each run of consecutive samples at or above threshold is one interval; filter on duration_seconds to drop short ones.
• detect_rapid_change(window=...) is deprecated and ignored.