How to identify hot and cold spots

A hot spot is the location that runs warmest, a cold spot the location that runs coldest. That sounds obvious, but 'warmest' can mean highest single reading, highest mean over the study, or the location that spends the longest time near the upper limit — and each definition can point at a different sensor.

Decide the criterion first

Write into the protocol which statistic identifies the spot — most protocols use the per-sensor maximum for the hot spot and per-sensor minimum for the cold spot, with the mean used to confirm that the location is consistently at the extreme rather than affected by a single event. Recording both prevents an argument after the fact.

Check the data before trusting the extremes

• A flat-lined sensor cannot be the cold spot.
• A single abnormal jump is more likely a sensor fault or a door event than a stable gradient.
• Sampling gaps near the extreme value weaken the evidence.
• Duplicated or non-monotonic timestamps usually mean an export or merge problem.

Then look at time

Plot the extreme sensors against time and against a reference sensor. A hot spot that only appears during defrost cycles or deliveries needs a different response from one that is warm all day. Excursions — periods outside the limits — should be listed with start, end, duration and peak, per sensor.

ValiTrac's analysis reports per-sensor minimum, maximum, mean, median and standard deviation, identifies hot and cold spots from the configured criterion, and lists excursions with duration and magnitude. Where sensor coordinates are not provided it says so rather than inventing a spatial map.