Commercial weather APIs usually consolidate information most governments provide for free, for the benefit of nautical/aeronautical trade. If you just need local data, search for you government's API.
Not an API, but I really like supercell wx for a local radar terminal. Cross platform and works great on Linux. It does require an API key to do the map rendering, but the actual weather data is pulled from government sources (free)
Yep, everyone I know in Australia just uses the BOM (the government bureau of meteorology) app. (If they aren't just using whatever build in weather app they have, that is.)
I've never understood what benefit any other app could provide given they are just ingesting the BOM data.
and for a streamlined way to grab weather forecast data directly from multiple governments national weather service APIs, there’s the library UniWeather.js:
On the app side, what's the best (global) app for rain forecasting over the next few hours (paid or not)?
Here in the Netherlands everyone uses "buienradar" which is limited to the Netherlands, has very bad privacy, and is also not super great at predicting rainfall.
It depends where you are in the world. The general solution involve an app that provides two types of forecast products: (1) a short-range, high-resolution numerical weather forecast over your country / domain of interest, refreshed rapidly (about an hour or so) and providing forecasts out to about 24 hours; and (2) a radar-based nowcast which extrapolates very short-term (~2 hours out) forecasts solely for rain.
The limitations straightforward. For (1), very few countries have access to such a forecast system outside of the US and continental Europe, and virtually no private company runs comparable systems (at least in the B2C space). For (2), very few countries have high-quality doppler radar networks and make the output available for these applications.
There really isn't a one-size-fits-all solution to this problem, despite what the umpteen-gazillion weather apps on the Play Store or Apple Store will try to sell you.
OWM constantly shows the high temp 10 degrees higher than actual on warm days. I emailed them.. 2 years ago? to notify them of this, they said they know... It's still a problem. Santa Clara, CA.
How do apps like Carrot get "hyper local" radar maps? Like ive never understood how these apps like Carrot have like 5-6 radar types if they are just calling someone elses API.. Id have to imagine theres some very heavy vector like DBs of points of xy radar raw data? Then they visualize it?
I'd like to see a global foundation managing open standards, with local operators (commercial, nonprofit, academic, government) running radars, and multi-layered funding (enterprise subscriptions, public-good subsidies, and co-op models) sustaining the system. It becomes the “public utility of weather data”, much like the internet itself.
Kudos to the dev for offering users alternatives. Tomorrow.io in the Carrot app is pretty solid for the most part, happy to see it make the list.
Dark Sky was one of the absolute best weather apps I've ever had the pleasure of using. I have no idea how Apple managed to mangle it to the current state it's in with their weather app.
On the topic of weather radar: I've been wondering what it'll take for Apple (arguably one of the most widely consumed weather services at this point!) to finally consume radar data available under open data licenses.
Some countries only make their data available commercially, but e.g. Germany does make high-resolution radar available under CC-BY 4.0, yet Apple seems to be completely ignoring all of it.
16 comments
[ 3.7 ms ] story [ 47.7 ms ] threadHere's the National Weather Service API, for those in the US: https://api.weather.gov
Not an API, but I really like supercell wx for a local radar terminal. Cross platform and works great on Linux. It does require an API key to do the map rendering, but the actual weather data is pulled from government sources (free)
https://github.com/dpaulat/supercell-wx
I've never understood what benefit any other app could provide given they are just ingesting the BOM data.
https://github.com/ghobs91/UniWeather.js
Here in the Netherlands everyone uses "buienradar" which is limited to the Netherlands, has very bad privacy, and is also not super great at predicting rainfall.
I also use RadarScope but that's more to see the intensity of nearby rain cells and try to guess for myself their movement and evolution
https://www.emc.ncep.noaa.gov/emc/pages/numerical_forecast_s...
Updated four times per day and has predictions out to about two weeks. It's used as the core input of most weather forecasts.
The limitations straightforward. For (1), very few countries have access to such a forecast system outside of the US and continental Europe, and virtually no private company runs comparable systems (at least in the B2C space). For (2), very few countries have high-quality doppler radar networks and make the output available for these applications.
There really isn't a one-size-fits-all solution to this problem, despite what the umpteen-gazillion weather apps on the Play Store or Apple Store will try to sell you.
https://apps.apple.com/in/app/drops-with-rain-radar/id981543...
Dark Sky was one of the absolute best weather apps I've ever had the pleasure of using. I have no idea how Apple managed to mangle it to the current state it's in with their weather app.
Some countries only make their data available commercially, but e.g. Germany does make high-resolution radar available under CC-BY 4.0, yet Apple seems to be completely ignoring all of it.