Overview
Time series data has a property that cross-sectional data doesn't: temporal order matters. A cleaning decision that is valid for a static dataset can destroy the temporal structure that makes time series analysis meaningful. Forward-filling a gap in a sensor reading is valid. Forward-filling a gap in a stock price creates look-ahead bias. Linear interpolation across a known system downtime produces phantom data points that never existed.
Time series cleaning requires understanding the data-generating process: why is the gap there, what should the value have been, and what does the downstream analysis assume about the data's temporal properties. The answers determine whether to interpolate, forward-fill, flag, or exclude.
The Time Series Data Cleaning & Gap Handling Prompt generates a complete temporal cleaning specification: gap classification by cause, interpolation method selection by gap type and series characteristics, temporal outlier detection, resampling strategy, and a validation framework that tests whether cleaning preserved temporal structure.
What you get: - Gap classification by cause (missing vs. zero vs. system downtime) - Interpolation method selection by gap length and series characteristics - Temporal outlier detection distinguishing spikes from level shifts - Resampling strategy for irregular to regular frequency conversion - Temporal structure validation post-cleaning
Built for: data scientists, ML engineers, and analysts working with sensor data, financial time series, event logs, and any temporally ordered data.