Two measurements taken at different times may tell a false story
A temperature sensor reports every minute while another reports every fifteen minutes. Comparing rows by position can create misleading differences. Time-series alignment ensures observations are compared using meaningful time references.
Define the analysis question
Real-time monitoring, hourly averages and long-term trends may require different alignment methods.
Preserve observation timestamps
The time a value was measured differs from when it was received or published.
Handle different sampling intervals
Aggregation or interpolation may be appropriate in some analyses, but assumptions must be explicit.
Respect missing data
A gap should not silently become zero or an invented observation.
Account for clock drift
Unsynchronised devices can create apparent leads or lags.
Document transformations
Users should know how aligned values were produced and which original observations contributed.
Break-test: a delayed batch arrives after the comparison was published
Can the system revise the series transparently without misrepresenting the original timing?
Connected knowledge
See Clock Synchronisation, Observation Timestamp Integrity and NEA.