Which default Weather service in /e/?

Here in Canada, where the weather is considered a national pastime, the federal government weather service provides forecasts and climate information. They have just released a lovely mobile app WeatherCAN.

The forecasts are reasonably good although, in my view, the Weather Underground (WU) forecasts are better. WU also has global coverage. Unfortunately, WU which started off as a grass roots movement was taken over a few years ago by the Weather Channel and its digital service was then spun off to IBM in 2016. I say unfortunately because my sense is that they are reducing support to the WU to focus on their flagship products The Weather Channel app and its web manifestation https://weather.com.

On the other hand, as corporations go, IBM is relatively more privacy-centric than the #axis-of-digital-privacy-evil heavyweights, Google, Facebook and Amazon . Perhaps it is because IBM has primarily served the entreprise sector where privacy and security receive more emphasis than in the consumer sector.

As I said, we Canadians are mildly obssessed with the weather :smirk: !

1 Like

To me, it tells wrong location it is inexactness

https://weather.com do not have weather data for all cities.
The 10 cities around me aren’t recognized

I just use the web :slight_smile: although the weather widget is nice.

Aucune difficulté à trouver ceux-ci…

In my location, the current temperature and predictions are totally wrong and thus useless. (OpenMap under /e/ version 0.5).

I suspect it may not be looking at the correct location. It shows the correct city name, but since I can’t see the state or country or coordinates, I don’t really know what the selected location is. I’ve been unable to get believable results using either location services or manual selection.

The weather on the OpenWeather website is much better, but still not terribly accurate.

Does NOAA (US Weather Service) have usable APIs? I suspect they are good on privacy, but not global data.

Is there a way to just turn off or disable the weather app?

Yeah… OpenWeatherMaps is nice in general but it gives real shitty forecasts for e.g. Sweden and sometimes in Germany, too… I’m using yr.no (which is a service of met.no) mostly and already thought about making a Lineage OS weather provider using data from yr.no. The thing is that they say there will be a fresh API soon, so I did not have enough motivation to mess around with the old XML API.


Source: https://hjelp.yr.no/hc/en-us/articles/360009342833-XML-weather-forecasts

But as I guess yr.no might not give quite perfect data for e.g. southern America, it might be good to stay with using Lineage OS weather provider API, so it is possible for everyone to choose their own weather provider (ofc there’s not a weather provider API for every source, yet - but maybe we could try to make it as easy as possible to create new ones?)

@Manoj @gael

I think that the creation of a survey would make it possible to have more visibility on the returns of the community by selecting the services mentioned in this subject.

We could ask all the users who wish to vote, to test the location and the forecast of the different services on the internet before voting and why not leave two choices instead of one.

1 Like

I did try openweathermap several times, but still coming back to https://www.YR.NO/en/

openweathermap is not able to find forecast in my village, always point me to closest big city which is cca 60km away {I live in EU}.

yr.no has not problem to show a forecast for my village, I did compare it with national service and it is quite close maybe in some situation better.

I just tried Klara and it seems to work very well!

Edit: I just see that Klara uses https://www.yr.no/.

Edit 2: An API is available and usable with latitude and longitude, for exemple: https://api.met.no/weatherapi/locationforecast/1.9/?lat=46.9558&lon=-1.35223. So, for developers, something can be done. :grinning:

Edit 3: If you want some code, Weather Forecast from F-Droid uses yr.no API.

1 Like

I think we need to beware of privacy-invasive (location-invasive) apps…
For me, web-based instead of a widget is priority because so many apps (weather.com :-1: , and others) spy so badly!

2 Likes

My location “München, DE” seems to exist in the database several times. Typing “münchen” presents “Munich, DE”, “München, DE”, “München, DE”, “München, DE”, and “München, AT”.

“Munich, DE” as first suggestion seems to be the right one.

“München, DE” as second suggestion seems to be a town called “Tannroda, DE” (380 km distance to Munich, DE)

“München, DE” as third suggestion seems to be a town called “Rupprechtstein, DE” (199 km distance to Munich, DE)

“München, DE” as fourth suggestion seems to be a town called “Bomsdorf, DE” (622 km distance to Munich, DE)

“München, AT” as fifths suggestion seems to be a town called “Prag, DE” (193 km distance to Munich, DE)

That alone is pretty strange except the first suggestion “Munich, DE”.

Anyway, I have no complaints about accuracy (if the location is right :))

But I have a missing feature about the weather app nobody seems to miss: The feature to save locations (at least 10). Does someone know if this feature exists and I’m just unable to find it?

Thanks!

1 Like

Аgree with you! A few days ago, I’ve been examining all the up-to-date weather apps I could find in the F-Droid. Forecastie appeared to be the best among them.

2 Likes

Interesting observation:

Two weather apps Forecastie and Weather
Both are FOSS and available from F-Droid Store
Both offer hourly forecasts (in contrast to Good Weather)
Both rely on OpenWeatherMap
Both predict the weather DIFFERENTLY (!)

Same for me. I am not localized. Since I changed London for Paris, the app automatically switch to Saint- Merri.

My impression has been that the IBM weather services, weather.com, The Weather Channel Android app and Weather Underground are all relatively non-invasive. For example, the only permissions the Android app request is location which it needs in order to provide you location-centric weather information. What is your evidence for the assertion that they are violating privacy?