We already can get hyper local weather data: every airport (and there are a lot of small airports in the U.S.) and tons of public buildings have weather stations on them. I guarantee you can get weather data within a few miles easily.

As to everything else: no, that would be creepy and the post office shouldn't do it. They don't need to become profitable or do off-mission things just to please the people that think it should be run like a business for some reason even though the whole point of the post office is to actually go everywhere (even where it wouldn't be profitable for FedEx et al to go) and be cheap (ie. not to be run like a business, to be run like a public service, which is what it is).

I said the same below but it applies here too: Google already drives around a fleet of vehicles to collect this data. Wouldn't it be better for them to give all that money to the postal service to subsidize mail delivery while also getting cars off the road?

And I'm talking about truly hyperlocal data. When I load up Weather Underground or Purple Air, there are a bunch of sensors near me, but the closest is one mile away. I know the route of my postal carrier. He starts his day less than one mile and stays within one mile of my house his entire day. I'd love to have daily temp and air pollution readings from sensors that are near my house all day.

There aren't that many cars that it would get off the road and it's not on mission and outside the realm of their expertise, so no, it's not a good way for them to make money which they shouldn't be doing anyways.

This weather data wouldn't be as useful as you think it would be. You'd be getting data only taken once per day more or less at the same time. You're much better using the airstrip a mile away that has a sensor that takes readings every 10 minutes or whatever.

There are 190,000 mail trucks delivering mail every day. I’ll bet NOAA would find 190,000 mobile sensors collecting data eight hours a day, most within a very small radius, quite a compelling dataset. There aren’t that many airstrips. Not even close.

And who cares if it’s not on mission? We can change the mission.

That's a fair point, NOAA might have a use for it, I don't know.

However, selling data to Google still seems like a bad idea. Things that are off mission and the point is to make money are just a distraction with no good purpose that's going to make the post office be greedy instead of just focusing on delivering the mail.

I could maybe agree that there was some value in having a public set of streetview like images (which is the more likely thing they'd do, not sell it to Google or whomever which would just be terrible), but it also seems deeply creepy. Then again, Google is already doing it so I suppose there's an argument that at least this way it's public domain.

Why don't you buy a personal weather station and connect it to wu? This seems more sustainable than lugging around that equipment on a truck, which increases the fuel cost. Plus, what are the chances of having a USPS vehicle within a mile of you at any given time? It's 6am on a Sunday and I want to know the weather.

I don't want the data for me. I want people who study long term weather patterns to have detailed hyperlocal datasets from 190,000 daily collectors.

I suppose google and facebook selling your data is not enough for you...

Google already drives around a fleet of vehicles to collect this data. Wouldn't it be better for them to give all that money to the postal service to subsidize mail delivery while also getting cars off the road?

I'd rather have the government build a public domain dataset as SamWhited suggested (and equally available to Google's competitors, projects like OSM, and any creative HNer) than have them collect proprietary data for Google and help lock in Google's competitive advantage.

Because everyone knows how good the government is with our data