Why the app doesn't need to measure constantly

Air pressure changes over hours. That single fact is what makes the whole project technically possible — and the battery lasts the day.

Aneroid measures air pressure every five minutes while the app is open, and less often when it isn't. That sounds like an admission — an app that doesn't measure constantly must surely miss something. It doesn't miss anything. Air pressure is one of the slowest phenomena in the weather, and that's not a claim from the app. It's the standard meteorologists have used for decades.

Three hours, not three minutes

The US National Weather Service, NOAA, works with a concept called pressure tendency. It has two parts: the pressure change itself — the net difference between the reading at the start and end of a period — and the pressure characteristic, meaning how the pressure has moved along the way, for example falling and then rising again. That period is usually the three hours before an observation. Not three minutes. Not thirty seconds.

The pattern over that period is recorded in three categories: falling, rising or steady. It isn't a fine-grained curve with hundreds of points. It's a coarse assessment of which way things are going, set against a period long enough for a real weather change to show itself. It's the same scale an app can lean on.

When the phenomenon itself plays out over hours, there's no reason to sample it in seconds. A reading every five minutes catches an entire weather change several times over while it's still happening.

The number you can't trust — and the one you can

There's another point that makes the matter even simpler. A study in the Journal of Atmospheric and Oceanic Technology put a number on how much you can actually rely on a smartphone barometer. The conclusion split the number in two. The exact pressure value isn't reliable — phones sit at different heights and have different sensor bias, so two phones side by side rarely show the same figure. But the tendency — whether pressure is rising or falling — is extremely reliable.

That's exactly the tendency Aneroid needs. Not the absolute figure, but the direction and speed over three hours. And that can be read off with far fewer measurements than you'd think. Two phones lying side by side on the same table can easily show two different pressure readings. They will almost never show two different directions once the pressure is genuinely moving.

That's also why an app doesn't need to conjure up a perfectly calibrated barometer to be useful. It just has to follow a curve over hours — not reproduce a figure another phone would disagree with.

The risk isn't the sensor

The barometer itself is a solved problem. It has been in the iPhone since model 6, and what the app actually needs from it — a tendency over hours, not a stream of second-by-second figures — is cheap to collect. The real technical uncertainty lies elsewhere: whether a phone sitting in a pocket for days manages to gather enough readings in the background to draw that curve at all. And whether the app is still installed for long enough. Studies of health apps in general show that around 71% of users are gone again within 90 days — which means that even technically perfect background logging is pointless if no one is there to fill it in.

The same literature review points to part of the explanation: studies of apps without feedback, reminders or any kind of support more often cite exactly that as the reason users disappear. An app that just logs quietly in the background and rarely asks for anything can be forgotten just as easily as one that spams notifications — just in a different way.

The sensor was never in doubt. Whether anyone would leave it switched on was.