Should DIY air monitors show their uncertainty?

I’m building a Raspberry Pi air-quality monitor with inexpensive particulate and environmental sensors, and I’m stuck on the display question: should every reading show calibration age, estimated sensor drift, and an uncertainty range, or is that too much honesty for a household gadget?

The EPA treats low-cost sensors as useful for supplemental, informational monitoring, but readings can be affected by humidity, temperature, pollutant mixtures, and limited selectivity. A clean “PM2.5: 12” looks wonderfully finished, yet may imply a confidence the hardware and correction model don’t deserve.

I’m leaning toward a simple e-ink number for quick glances, with the calibration date, drift estimate, and uncertainty band on a local dashboard. That keeps the device usable without hiding the caveats. Do you expose uncertainty in your own projects, or do you think this makes practical devices harder to trust? Examples and counterarguments welcome.

I’d expose the metadata, but not necessarily beside every number. “12 ± estimated range” is only meaningful if the range reflects the sensor, calibration method, and conditions; it is not a universal error bar. A dashboard tab or tap-through view seems more defensible than pretending one formula covers every installation.

The e-ink/dashboard split sounds right. Most people need a trend or an alert, not a lecture from the screen. But calibration age should be visible in the same place as the reading, even if it’s a small status line. Hidden caveats tend to become forgotten caveats.

I’d also show a stale-data indicator. Drift and uncertainty are not the only failure modes; a stuck value can look perfectly plausible. Keep the primary display simple, but make status, last calibration, and sensor health first-class fields in the local data model rather than decorative text.

Counterargument: displaying a guessed range may create false precision of a different kind. If the correction model has not been validated against a reference monitor, “±8” can sound more scientific than the raw 12. I’d show calibration age and a qualitative confidence state first, then add numeric uncertainty once the method earns it.