Catch up on stories from the past week (and beyond) at the Slashdot story archive

 


Forgot your password?
Close
typodupeerror
×
IT Technology

The Problem With Weather Apps (theatlantic.com) 57

Posted by msmash from the closer-look dept.
An anonymous reader shares a report:Weather apps are not all the same. There are tens of thousands of them, from the simply designed Apple Weather to the expensive, complex, data-rich Windy.App. But all of these forecasts are working off of similar data, which are pulled from places such as the National Oceanic and Atmospheric Administration (NOAA) and the European Centre for Medium-Range Weather Forecasts. Traditional meteorologists interpret these models based on their training as well as their gut instinct and past regional weather patterns, and different weather apps and services tend to use their own secret sauce of algorithms to divine their predictions. On an average day, you're probably going to see a similar forecast from app to app and on television. But when it comes to how people feel about weather apps, these edge cases -- which usually take place during severe weather events -- are what stick in a person's mind. "Eighty percent of the year, a weather app is going to work fine," Matt Lanza, a forecaster who runs Houston's Space City Weather, told me. "But it's that 20 percent where people get burned that's a problem."

No people on the planet have a more tortured and conflicted relationship with weather apps than those who interpret forecasting models for a living. "My wife is married to a meteorologist, and she will straight up question me if her favorite weather app says something different than my forecast," Lanza told me. "That's how ingrained these services have become in most peoples' lives." The basic issue with weather apps, he argues, is that many of them remove a crucial component of a good, reliable forecast: a human interpreter who can relay caveats about models or offer a range of outcomes instead of a definitive forecast. [...] What people seem to be looking for in a weather app is something they can justify blindly trusting and letting into their lives -- after all, it's often the first thing you check when you roll over in bed in the morning. According to the 56,400 ratings of Carrot in Apple's App Store, its die-hard fans find the app entertaining and even endearing. "Love my psychotic, yet surprisingly accurate weather app," one five-star review reads. Although many people need reliable forecasting, true loyalty comes from a weather app that makes people feel good when they open it.
This discussion has been archived. No new comments can be posted.

The Problem With Weather Apps More

The Problem With Weather Apps

Comments Filter:

  • You could get close to that by just giving the previous days forecast or in the sahara "Hot and dry" will probably hit 99.999%

    • Re: (Score:3)

      by pruss ( 246395 )

      90% at five days out and 80% at seven days ( https://scijinks.gov/forecast-... [scijinks.gov] ) is pretty good, I think. I doubt you are going to get that good by saying it'll be the same weather in a week as yesterday, except in places like the Sahara.

      • Could work in Florida for most of the year. Hot, humid, chance of rain.

        • Re: (Score:3)

          by dbialac ( 320955 )
          Well May to October, kind of. Along the coast, high around 90 with isolated thunderstorms starting around 1, weather then cooling off by evening. During the rainy season, the low temperature for the day is during the thunderstorm, not at night.
        • Our (non-US) weather service has taken this to heart, they love reporting "chance of rain" because you can never be wrong with that. There's a chance of rain in the Sahara desert, just not a very big chance. Now if they could tell us what the chance is, and how much rain, that'd be an improvement, but all they'll ever say is "chance of rain". Our entire multimillion-taxpayer-dollar-funded weather service could actually be replaced by a parrot trained to say "chance of rain, chance of wind, more severe in

    • Oddly enough, the MSN weather page has been nearly 100% accurate (at least with temperature) in my experience. That's all I really care about in a weather app, I don't care so much about wind, precipitation or other factors.

    • Re: (Score:2)

      by dbialac ( 320955 )
      Since TWC switched over to using IBM's computerized systems, their forecasts are abysmal for the area I live in. Abysmal as in they can't even get the current temperature right for the city I live in. They now have their models providing "current radar", which can be unbelievably off. On multiple occasions I've seen the generated image show rain where there was none and worse, no rain when there was a torrential storm. The latter is flat out negligence and I'm surprised they haven't been dragged to court ye
      • If you think that's bad, look at how much they've fscked up Weather Underground since IBM took over.

        • If you think that's bad, look at how much they've fscked up Weather Underground since IBM took over.

          Thank you for this. I worked on the original Weather Underground apps and I'm sad every time I have to use the new one.

  • TL;DR (Score:3)

    by ddtmm ( 549094 ) on Tuesday April 11, 2023 @12:36PM (#63441670)

    Although many people need reliable forecasting, true loyalty comes from a weather app that makes people feel good when they open it.

    • Re:TL;DR (Score:4, Interesting)

      by jenningsthecat ( 1525947 ) on Tuesday April 11, 2023 @12:57PM (#63441756)

      Although many people need reliable forecasting, true loyalty comes from a weather app that makes people feel good when they open it.

      I also came here ready to paste that quote. I was going to add that a weather app which predicts intermittent unicorn showers has a chance of owning the internet.

    • There's actually a really simple way to find a good weather app, take the top half-dozen or so that you like, record their predictions for a week or two, and compare them to the actual weather that turned up. Then take the one that gets it right the most often. Doesn't matter whether it's getting its results from NOAA data or by reading sheep entrails, if it's more accurate than all the others it's fine.
  • We pay taxes for all that weather data, and thus why wouldn't private companies profit from it? Last Week Tonight did a funny expose on it: https://www.youtube.com/watch?... [youtube.com]

    • Re: (Score:2)

      by ranton ( 36917 )

      We pay taxes for all that weather data, and thus why wouldn't private companies profit from it? Last Week Tonight did a funny expose on it: https://www.youtube.com/watch?... [youtube.com]

      What weather data are you paying extra for? The weather app on my phone is free. I guess there are some ads if I want to visit their website for radar projections, but nothing obtrusive.

      • Re: (Score:1)

        by Anonymous Coward

        are you serious? NOAA has to be funded somehow, and no those ads on your weather app of choice are NOT how that is done.

  • Weather Underground. (Score:5, Informative)

    by Major_Disorder ( 5019363 ) on Tuesday April 11, 2023 @12:45PM (#63441702)
    It was much better before IBM bought them, but I still find that it is VERY accurate.

    • Re:Weather Underground. (Score:5, Informative)

      by turp182 ( 1020263 ) on Tuesday April 11, 2023 @01:09PM (#63441790) Journal

      Same here, although I keep Accuweather around as I prefer it's radar presentation (why the f are weather radar image streams so crappily presented/performant???).

      • I refer to the site all the time via a browser. I'm leery of weather apps in terms of data collection, but I'd happily pay for one that I knew wasn't hoovering up personal data. (I usually say no to all unnecessary permissions but lately articles indicate that android apps have some backdoor options to still collect despite permissions)

        I think the real problem is figuring out which weather apps are malware and spyware. (probably most of them.) Weather is such an easy service to provide thanks to NOAA and

    • WeatherUnderground took its first degrading in quality when The Weather Channel bought them. Then IBM bought the data services of The Weather Channel (leaving the cable studios to some other company). IBM's weather.com is still one of the better out there, though, and their app is private-labeled by many TV stations around the country.

      • Is anybody actually using personal weather stations to improve models? I live in an area with huge microclimates, and it is funny that we can get an inch of rain while the Apple app indicates zero rain in the previous 24 hours.

        I guess the only way to care about microclimates is if it is commercially meaningful locally, but I wouldn't think it would be that hard to customize.

        • Re: (Score:2)

          by erice ( 13380 )

          Is anybody actually using personal weather stations to improve models?

          Weather Underground does this and, here in the SF Bay Area, it makes a huge difference compared most apps, like Accuweather that rely on airports. Unfortunately, data from personal weather stations can be wildly inaccurate. The forecast may be much more localized but it is still wrong.

        • Re: (Score:2)

          by nasch ( 598556 )

          I live in an area with huge microclimates

          Sounds like a contradiction in terms.

    • Re: (Score:2)

      by dbialac ( 320955 )
      It's so accurate that it can't give me an accurate current temperature where I live.

    • Re: (Score:2)

      by Dadoo ( 899435 )

      I came here to say "I just go to Google and type 'my_city weather' and I don't even have to hit enter to get the temperature", but yours is better.

  • It can't be that hard to set an ML engine like GPT or something else to scan data and images and correlate the data taken in with actual results.

    Do this for 2-3 years for a given region under continuous weather satellite and sensor surveillance and the AI model will predict the weather notably more accurate than any human can. And we likely won't know how and by what criteria.

    Point in case: I recently heard from a doctor that they trained AI on x-rays. The AI is 86% accurate on its own. Humans are like 75%

    • This is why IBM purchased weather.con

    • ML is used widely for these apps. They are mainly stripping bias from NCEP/ECMWF grids and downscaling from those grids to points. Easy and effective, most of the time. The edge cases (extreme events) are where details in the forecast really matter, and edge cases are where ML is notoriously weak in going outside training. This is why the NWS local offices have not gone away, because experts matter most when lives and property are on the line.

    • Re: (Score:2)

      by edwdig ( 47888 )

      You've got two problems here.

      First is the AI really only learns to recognize things it's seen a lot of, so it won't recognize the really bad events that we care most about. It'll be great at telling you about your average days, and it won't know what to think about a really bad hurricane.

      Second, the climate has been rapidly changing. Previously rare extreme events are becoming much more common. General severity of the bad storms is getting worse. Training an AI off past trends is going to result in you cons

  • Use multiple weather apps. I'm surprised there's no aggregator service. In the meantime, I get three or more weather reports and use my judgement. They almost always disagree.

    As for a weather app making you "feel good?" WTAF. People with a data fetish?

    • There are aggregators at the pro level (e.g. NAEFS and TIGGE). Disagreement is good provided verification falls in that distribution at the correct rate. E.g., if you have a reliable forecast distribution, it better rain 10% of the time when you put out a 10% probability of precipitation, no matter how it makes people whine that forecasts are "unreliable." That's just statistical ignorance.

    • As an amateur astronomer, I appreciated 7Timer [7timer.info] which offers astronomical seeing enough to make a free iOS client for it (called Xasteria [apple.com]) and donate the servers of the project.
      However, I also live in England, where the clear sky opportunities are few and a single weather source is not reliable enough. My solution is to check weather sources that use different models: 7Timer uses the NOAA hydrostatic model, add to that something like YR.no [www.yr.no] for the non-hydrostatic ECMWF and also something that has good short-

    • This reminds me of Segal's Law:

      A man with a watch can tell you what time it is. A man with two watches, is never quite sure.

  • "My wife is married to a meteorologist, and she will straight up question me if her favorite weather app says something different than my forecast," Lanza told me

    Well Lanza, maybe you should make an app for your weather predictions so your wife can put it on her phone. Problem solved, trust restored, the app says it so it must be true.

  • National Weather Service publishes their own forecast - it's just a page I open on my PC or phone - but I figure it's the best since they originate the data.

    • Re: (Score:3)

      by g01d4 ( 888748 )

      National Weather Service publishes their own forecast

      Called the Area Forecast Discussion. Basically they discuss the model forecasts and establish confidence based on the amount of agreement. If the forecasts don't agree then they use their experience to apply a weight to forecasts relative to their past success in the same instances. So, yes, 'AI' can probably step in on this unless it can be shown there's some magic intuition that improves the predictions.

  • ... there seems to be a bit of a trend towards hiring on-air weather personalities instead of meteorologists to provide the weather forecasts in the TV news (one local stations describes them as a "weather presenter"). These on-air weather personalities do little more than read the visuals that the weather models present, and provide no insight behind what the models are showing and do not comment upon the accuracy of what the models might be showing.
    • A few days ago I was viewing one of the weather channels. When the presenter looked at one Doppler radar image, he launched into open prayer for people in that area (maybe Arkansas or Mississippi).

  • Predicting the weather is hard (Score:4, Interesting)

    by Retired Chemist ( 5039029 ) on Tuesday April 11, 2023 @01:20PM (#63441842)
    Predicting the weather is hard. Their many computer models and it is well known that none of them are truly accurate. One of the primary skills of a modern meteorologist is knowing when to ignore a particular model's predications. That said, it does not appear to me that the quality of the predictions has improved very much from forty years ago, when it was done by humans looking at weather maps without the aid of any fancy computer models. When it comes to solving some kinds of very difficult problems where chaotic conditions occur, humans are still sometimes better than any algorithm. In any case, since the National Weather Service provides predictions for free, why do we need any apps at all?

    • Re: (Score:2)

      by nasch ( 598556 )

      In any case, since the National Weather Service provides predictions for free, why do we need any apps at all?

      If you're satisfied with the NWS web site, you don't.

    • Re: (Score:2)

      by pjt33 ( 739471 )

      In any case, since the National Weather Service provides predictions for free, why do we need any apps at all?

      The weather app I have installed on my phone is from the national weather service of the country where I live. It's far more convenient than navigating a website every time I want to check the weather for the city where I live.

  • It's fixed everything else in my life. Why not the weather?


    ...and then there was one.

    • As an AI language model, I do not have access to meteorological instruments with which to measure the atmosphere or make predictions. However, you would do well to just look out the window once in a while, or pull up a radar map to view any active storm systems heading your way. Your grandma can also tell you about incoming storms in advance because the drop in barometric pressure causes her arthritis to ache.

      It is important to remember that weather forecasting is largely guesswork, and those people are

    • ChatGPT data is limited to 2021 and older. https://blog.gitnux.com/chat-g... [gitnux.com]. It could probably predict weather for one of those years!

  • The problem with weather apps in Brazil is that they mistakenly use the word "clima" (climate) instead of "tempo" (weather), just because "tempo" also means "time".
  • One of my local TV stations (WTOL in Toledo) pushes their weather app hard. They mention it at least 5 times during every weather segment. Why? Ads, just because they get money for pushing ads. So transparent. It's obvious these poor weather men/women are being used to push the advertising agenda. I tried their app. I don't want to wait for a full page pizza ad to disappear first and then ignore all the ads scattered around just to get a quick peek at the forecast. Yuck. Deleted.

  • I own a weather station, which is showing me very nice local data, including PM10 and PM2.5 values. Pretty horrific, those values...

  • App? No one needs a weather app. (Score:5, Informative)

    by geekprime ( 969454 ) on Tuesday April 11, 2023 @03:25PM (#63442238)
    In the US https://www.weather.gov/ [weather.gov] will let you enter your zip+4 or street address or even gps coordinates and show you not only your forecast but radar on demand and historical weather information, and there is not a single ad on the entire site. It's SO good that the owner of a certain weather site/app has bribed multiple senators multiple times to try to get it taken off the internet because it interferes with his business model. https://stormeyes.org/wp/2005/... [stormeyes.org]

  • https://nwsnow.net/ [nwsnow.net]

    It is a well done app of the National Weather Service data and human forecast. It avoids the Google-plex, and I have used it for years very successfully. Oh yeah, buy the dev a coffee. Super cheap; super sustainable.

  • I paid for the the Met Office app (United Kingdom) as I spend 99.5% of my time here these days. It does include a video weather briefing - and in those 2-3 minutes I get told a story - one that is often accurate in its prognostications. So therefore I'm inclined to believe there is value to a human interlocutor to the data - someone that gets it a deeper level than my Dunning-Kruger level meteorology. Their app is not fancy - no animations etc. Just some data. Then a video. But I like it.

  • High chance of sunshine and overall great weather. If by chance it does rain, GREAT!!!

  • I bounce between apps (web and iOS) depending on what I'm looking for. I'm usually interested in the week's outlook for bike rides.

    Weather Underground has the best chart for the 10 day forecast. You can see temp, wind speed and precipitation forecast in one easy to read chart.

    For "now" the updated iOS and Ventura weather is a great improvement as you can see the state of everything in one page. Charts are good except you can only see one measurement at a time so it's bouncing between temp, wind and now with spring upon us, the UV index for the day. On my Apple Watch I keep a screen with complications for weather, wind and UV but they're tiny. When riding I use the large single complication for wind. Once on the bike that's all that really matters and can vary within the scope of a 2-3 hour ride.

    Windy is nice when exploring broad geographies for curiosity or newsworthy weather events.

  • I think the weather apps are fine as far as they go, but they tend to present data for a large area. I want them to take account of my specific location and present results based on that. If the rain is passing 10 miles north of me, don't tell me it is raining. I end up looking at the radar and judging for myself a lot of the time. There are results from weather stations that could be incorporated as well, I end up using Weather underground to see if it rained at a specific location, like if I want to visi

Slashdot Top Deals

The truth of a proposition has nothing to do with its credibility. And vice versa.

Close