THE RESULT
What the data shows
The largest computed metric is 0.9023 for nmhc_gt; the smallest is 0.03912 for c6h6_gt. Metric: missing_share (Missing share).
Entirely blank CSV rows and unnamed empty columns were removed; missing instrument readings are retained as NaN.
| field | observed | missing | missing share |
|---|---|---|---|
| nmhc_gt | 914 | 8,443 | 0.9023 |
| co_gt | 7,674 | 1,683 | 0.1799 |
| no2_gt | 7,715 | 1,642 | 0.1755 |
| nox_gt | 7,718 | 1,639 | 0.1752 |
| c6h6_gt | 8,991 | 366 | 0.03912 |
| pt08_s1_co | 8,991 | 366 | 0.03912 |
| pt08_s2_nmhc | 8,991 | 366 | 0.03912 |
| pt08_s3_nox | 8,991 | 366 | 0.03912 |
| pt08_s4_no2 | 8,991 | 366 | 0.03912 |
| pt08_s5_o3 | 8,991 | 366 | 0.03912 |
| t | 8,991 | 366 | 0.03912 |
| rh | 8,991 | 366 | 0.03912 |
THE METHOD
From source to answer
Field-level availability after converting the documented -200 missing sentinel.
The Python source exposes this study’s transformations. The complete project download includes shared preparation and evaluation routines.
THE NEXT DECISION
What follows from the finding
Prioritize recovery of reference measurements before calibrating sensors.
Where the conclusion stops
The -200 sentinel is treated as missing. Reference instruments have incomplete coverage; comparisons use paired observations. Sensor calibration is retrospective, not a health or regulatory compliance assessment.
Related studies may reuse observations or holdouts. These are historical analyses; associations and backtests do not demonstrate commercial impact. Further model tuning needs new, untouched evaluation data.