Skip to content
SkyLedger
All tools
Atmospheric Observation

Paired Sensor Bias Comparison

Compare synchronized sensor/reference pairs using signed bias, absolute error and a difference chart.

Use this tool

Preparing interactive controls…

Paste CSV or open a local file. Nothing is uploaded or stored. Limit: 1 MB and 1,000 data rows. Changes recalculate the report; keep a copy of your original data.

Pairs must be simultaneous, colocated, in the same unit and measuring the same quantity. A comparison is not a traceable calibration; no correction is automatically applied.

Analysis result

Pairs
3
Mean bias (sensor − reference)
1
Mean absolute error
1
Root mean squared difference
1
Maximum absolute difference
1
1, 12, 13, 1131.50.5Sensor − reference (input unit)Pair number
Lines join supplied points for readability; they do not establish continuous observations, including within gaps.
Input or flagged rows (all 3)
ReferenceSensorDifference
10111
20211
30311
Introduction & principle

For d=sensor−reference: bias=mean(d), MAE=mean(|d|), RMSD=√mean(d²). Units are whatever consistent unit the paired inputs use. Comparing colocated simultaneous values is not traceable calibration or a license to automatically correct a sensor.

Worked example

Reference 10, 20, 30 paired with sensor 11, 21, 31 gives bias +1, MAE 1 and RMSD 1 in the chosen unit.

Detailed usage guide
  1. 1Use reference,sensor CSV with at least two synchronized pairs in one consistent unit.
  2. 2Choose matched exposure and time intervals; do not compare different quantities or units.
  3. 3Read difference metrics and the chart, then investigate conditions before changing calibration.
Common problems & FAQ

Can I use the bias as a correction?

Not automatically. Bias depends on reference quality, exposure, time alignment and operating range. Investigate these before applying any adjustment; the report does not alter the sensor or source file.

Methods & references

Method reviewed November 4, 2026. These tools support learning and recording, not safety-critical decisions.

Related tools

Observation CSV Quality Check

Inspect a local observation CSV for missing values, duplicate or unordered UTC times and range flags.

Open

Observation Log Markdown Template

Prepare and download a Markdown observation sheet with observer, location and instrument notes.

Open