Author Topic: WU website down  (Read 4998 times)

0 Members and 1 Guest are viewing this topic.

Offline MrM1

  • North Central Florida
  • Contributor
  • ***
  • Posts: 141
    • Bethel Weather
Re: WU website down
« Reply #25 on: June 24, 2017, 04:31:44 PM »
Back up in the last 10 minutes here







Offline daman

  • Forecaster
  • *****
  • Posts: 520
Re: WU website down
« Reply #26 on: June 24, 2017, 04:41:21 PM »
Same here back up.
KMIBADAX38
AcuRite Iris/AcuRite Access

Offline jerryg

  • Forecaster
  • *****
  • Posts: 886
  • weather is never boring
    • victoria texas weather
Re: WU website down
« Reply #27 on: June 24, 2017, 04:46:41 PM »
wu back up and running

Offline jackman

  • Senior Contributor
  • ****
  • Posts: 251
    • Echo Pines WX Station
Re: WU website down
« Reply #28 on: June 24, 2017, 05:25:20 PM »
Yup, back UP and all my data is there too.

Offline openvista

  • Forecaster
  • *****
  • Posts: 481
    • marquetteweather.com
Re: WU website down
« Reply #29 on: June 24, 2017, 06:02:10 PM »
Yes, Wunderground has major uptime and quality control issues. It's very frustrating. But I have figured out how to grab forecast data from them using caching and backups where my site virtually never experiences disruption. It's possible. I didn't reinvent the wheel or anything.

Why do I continue using Wunderground? Simple. Forecast accuracy. I know from my own data collection, their forecasts are more accurate for my town than NWS. In fact, others (like forecastadvisor.com) have spent considerable time and effort tracking the accuracy of major forecast providers and WU consistently rates near the top, well above NWS (NDFD).

Hey, I'd love a weather API that was solid. But that certainly doesn't describe the NWS's new API!! That thing has been a total nightmare with blown deadline after blown deadline, incomplete and incorrect documentation, missing functionality, arbitrary data values, etc, etc. 

Forgive me for getting a little philosophical, but when you struggle to find someone who can screw a lightbulb in correctly these days, the problem might be a little larger than one particular organization. Quality, if it's even considered, is way down on the list of metrics used to evaluate the success or failure of most any venture now. Sad, but true. 
Davis Vantage Pro2 AC FARS | https://marquetteweather.com | EW7933

Offline Jáchym

  • Meteotemplate Developer
  • Forecaster
  • *****
  • Posts: 8605
    • Meteotemplate
Re: WU website down
« Reply #30 on: June 24, 2017, 06:30:32 PM »
Check DarkSky and YrNo, especially the latter one is much more accurate for my location compared to WU and they are both global as well

Offline openvista

  • Forecaster
  • *****
  • Posts: 481
    • marquetteweather.com
Re: WU website down
« Reply #31 on: June 24, 2017, 07:41:25 PM »
Check DarkSky and YrNo, especially the latter one is much more accurate for my location compared to WU and they are both global as well

Well, they may be "global", but Yr.no doesn't even attempt to present their data in English units. To me that says, they don't really care to serve U.S. visitors. Also, I've seen many forecast accuracy graphs of WxSIM users who compare Yr.no to WxSIM. It's bad when a well-funded government agency is badly outdone by $300 desktop software. In any event, I can't use them in the winter because, in my experience, they don't account for the Great Lakes in their forecasts. We get about 200cm of lake-effect snow here annually.

According to forecastadvisor.com, Dark Sky rates below even the NWS in accuracy (not just in my town, but generally). I know from my own observations they aren't close to WU's accuracy. I use them as a radar provider because they have a top-notch radar product. Precip modeling is the focus of their business. Temps not so much.

I know many here don't want to hear it, but Wunderground (WSI) is the most accurate forecaster, generally speaking, in the United States, based upon the available, empirical data. I'm absolutely sure there are exceptions, but can we move away from "they suck and I know it" toward "here's the data I've collected measuring their forecasted values against my highs and lows"? I'll show you mine if you show me yours.
Davis Vantage Pro2 AC FARS | https://marquetteweather.com | EW7933

Offline jackman

  • Senior Contributor
  • ****
  • Posts: 251
    • Echo Pines WX Station
Re: WU website down
« Reply #32 on: June 24, 2017, 08:16:33 PM »
...
I know many here don't want to hear it, but Wunderground (WSI) is the most accurate forecaster, generally speaking, in the United States, based upon the available, empirical data. I'm absolutely sure there are exceptions, but can we move away from "they suck and I know it" toward "here's the data I've collected measuring their forecasted values against my highs and lows"? I'll show you mine if you show me yours.

Well said thank you.

Offline Jáchym

  • Meteotemplate Developer
  • Forecaster
  • *****
  • Posts: 8605
    • Meteotemplate
Re: WU website down
« Reply #33 on: June 24, 2017, 08:52:51 PM »
I take your point, though I have to say that at least here it is quite the opposite, in my country (and I would say the whole Europe), yrNo is far more accurate than WU, in some cases WU is totally off, I mean like totally. Sometimes the difference is way over 5°C, which is almost 10°F. Especially more than 2 days ahead, I have very bad experience with it, and DarkSky is somewhere in the middle. The conclusion from this to me is that even though these services are all global, they probably work better in some parts of the world than in others. I cannot compare WU with NWS myself because obviously NWS does not provide forecasts for my location.

One thing I want to say though is that you cannot compare WxSim with pages like YrNo. Simply because WxSim uses data directly from your station so it has much more accurate input about your climate. Sometimes even just a few km/mi can make a huge difference and yrNo and other webpages like this depend on climate data for regions at a much larger resolution, not sure, but I think yrNo uses a grid of 15km. WxSim will get climate input which is spot on for your location. I dont have personal experience with WxSim, but Im sure it does a great job, Im just saying it wouldnt be fair to compare it with pages like yrNo.

Offline openvista

  • Forecaster
  • *****
  • Posts: 481
    • marquetteweather.com
Re: WU website down
« Reply #34 on: June 24, 2017, 10:56:17 PM »
WxSim will get climate input which is spot on for your location.

Oh you mean like this (found on WU's about page)?
« Last Edit: June 24, 2017, 11:19:14 PM by openvista »
Davis Vantage Pro2 AC FARS | https://marquetteweather.com | EW7933

Offline openvista

  • Forecaster
  • *****
  • Posts: 481
    • marquetteweather.com
Re: WU website down
« Reply #35 on: June 24, 2017, 10:58:09 PM »
Look, we can both argue about the effectiveness of such claims at localization. I've seen instances where WxSIM doesn't do so well. I've also seen instances where WU doesn't do so well. In fact, it wouldn't surprise me if, generally, WU was bad at international forecasting (although I'd still like to see the data). But how can you argue that it's unfair to expect a global forecast service to provide neighborhood weather? That IS Weather Underground's niche and reason for existence and why many of us continue to bear with them despite how often they disappoint in many other ways.
Davis Vantage Pro2 AC FARS | https://marquetteweather.com | EW7933

Offline Bashy

  • brecklandweather.com/meteo
  • Forecaster
  • *****
  • Posts: 1465
  • brecklandweather.com/index.php
    • Breckland Weather
Re: WU website down
« Reply #36 on: June 25, 2017, 12:29:42 AM »
WxSim will get climate input which is spot on for your location.

Oh you mean like this (found on WU's about page)?


Something tells me thats regarding Current Conditions, (i.e. hey, buy a station and you can see precise conditions
at your location), not forecast,

If you look hard enough, you start to see whats not there ;)

If there is one thing i have noticed and thats as soon as a weather underground posts is made and its about
something not right. Its always the same haters that reply ;) Just because WU didnt go right for some, does
not, now make it wrong for others.

Guys, ya hate WU, we get that, please, enough of the witch hunt, it really does not do yourself any favours
I doubt you are or will put anyone off using them, they can only find that out for themselves...

I cannot find another, on the spot, local stations current conditions lookup with the same or better UI, that also
allows you to send data and download it again anywhere else. I lost about 7 years worth of data, until i found
out i could import it easily from WU (thanks for that Jacym) and dont you dare say Awekas......

« Last Edit: June 25, 2017, 12:47:05 AM by Bashy »
Kind regards
Bashy

Offline openvista

  • Forecaster
  • *****
  • Posts: 481
    • marquetteweather.com
Re: WU website down
« Reply #37 on: June 25, 2017, 12:45:15 AM »
Again, from Wunderground's about section (https://www.wunderground.com/about/data):

Quote
All of our forecasts are generated from our proprietary forecasting system that leverages our vast amount of neighborhood weather data that we get from our community - we refer to the system as BestForecast™.

Our ever-expanding network of 250,000+ personal weather stations is the largest of its kind and provides us with a unique ability to provide the most local forecasts based on actual weather data points. BestForecast™ uses the most innovative forecast models available and cross-verifies their output with all of the localized data points. Only our unrivaled amount of local neighborhood weather data can generate forecasts for your front door.
Davis Vantage Pro2 AC FARS | https://marquetteweather.com | EW7933

Offline Bashy

  • brecklandweather.com/meteo
  • Forecaster
  • *****
  • Posts: 1465
  • brecklandweather.com/index.php
    • Breckland Weather
Re: WU website down
« Reply #38 on: June 25, 2017, 12:49:38 AM »
Thats a better quote, not so sure that that is anything for WU to boast about mind, HAVE YOU SEEN THE LOCALS DATA  :roll: #-o :lol:
Kind regards
Bashy

Offline openvista

  • Forecaster
  • *****
  • Posts: 481
    • marquetteweather.com
Re: WU website down
« Reply #39 on: June 25, 2017, 12:55:36 AM »
Thats a better quote, not so sure that that is anything for WU to boast about mind, HAVE YOU SEEN THE LOCALS DATA  :roll: #-o :lol:

As I said, your mileage may vary. But in my locale, they are better than the alternatives regarding temperature forecasts.
Davis Vantage Pro2 AC FARS | https://marquetteweather.com | EW7933

Offline openvista

  • Forecaster
  • *****
  • Posts: 481
    • marquetteweather.com
Re: WU website down
« Reply #40 on: June 25, 2017, 01:12:15 AM »
As far as the original quote (screenshot) vs the later one, if you look at the very first sentence of the screenshot it alludes to the conditions data improving their forecasts. Seriously, though, how long has WU marketed the use of PWS data in their BestForecasts? A decade? None of this should be news.
Davis Vantage Pro2 AC FARS | https://marquetteweather.com | EW7933

Offline Bashy

  • brecklandweather.com/meteo
  • Forecaster
  • *****
  • Posts: 1465
  • brecklandweather.com/index.php
    • Breckland Weather
Re: WU website down
« Reply #41 on: June 25, 2017, 01:30:54 AM »
I dont think it applies to me in the UK, cant do seen as their forecasts for us is not overly clever, i have nothing against WU.
Kind regards
Bashy

Offline gillesroszak

  • Contributor
  • ***
  • Posts: 114
  • Qui trop écoute la météo, passe sa vie au bistrot.
    • meteokerbach
Re: WU website down
« Reply #42 on: June 25, 2017, 06:03:50 AM »
that also
allows you to send data and download it again anywhere else. I lost about 7 years worth of data, until i found
out i could import it easily from WU (thanks for that Jacym)

+1  Thanks to Jachym

Offline ocala

  • Forecaster
  • *****
  • Posts: 4383
Re: WU website down
« Reply #43 on: June 25, 2017, 07:01:57 AM »
Interesting info about the accuracy of the the Best Forecasts versus the NWS Point Forecasts. I have never thought about comparing them.
I noticed for my backyard today the chance of precipitation is quite different.
Best Forecast is 80%
NWS Point Forecast is 30%
That's a rather large spread.
Going forward I'm definitely going to track the difference.

« Last Edit: June 25, 2017, 07:06:26 AM by ocala »
The blues had a baby and they named it Rock & Roll

Offline Jáchym

  • Meteotemplate Developer
  • Forecaster
  • *****
  • Posts: 8605
    • Meteotemplate
Re: WU website down
« Reply #44 on: June 25, 2017, 07:41:44 AM »
OK, so I just want to clarify why exactly Im so negative about WU...

There are things which are subjective - I dont like their page design, the navigation etc. However, I understand this is subjective and others might like it, so this is not something I see as negative, it just makes me less likely to use their page.

However, then there are absolutely black/white things they do wrong:

1. the way they actually use the PWS data - at our institute, we also use interpolation, i.e. you use nearby station´s data to make forecasts, you base it on several stations nearby. However the problem is that you can never do it in an environment such as WU (and in fact Awekas does this as well). Why? Simply because many of these amateur PWS are very badly sited and provide inaccurate data.
Let me give you an example:
You buy a Davis station, site it 2m above the ground above grass cover, make sure data is as accurate as possible. Both your neighbors find it "cool" to have a station as well, but they are no weather enthusiasts so they buy NetAtmo, mostly because it looks nice in their living room and place the outdoor sensor either next to the outside wall (which is of course heated in the winter) or in direct sunshine, without proper radiation shield. Both these stations now send inaccurate data, which could be quite similar, but different to yours. Now which station do you think will be seen as inaccurate based on the algorightms Awekas and WU use? Yes, of course your Davis, because you are a minority.
My point is, such interpolations and data quality checks can only be done if you have reference stations, which you know are accurate and you can only compare between those. Neither WU nor AWEKAS have this. And the number of people who use stations as more or less just fun, is larger than the number of people like us, I can assure you about that.

2. support
Im very disappointed with the support WU provides. I will again give an example of what happened several times in the past. Now, I know it can be difficult to keep things working 100% of the time. Of course there will never be a bug-free SW, or 100% reliable server. I understand that. However, several times in the past, WU was down, something was not working at all. And the only communication channel in that case is their FB page because the site itself was down.
In an ideal situation, they would post something like "We know about the problem, we are working on it, we expect it to be fixed in.... hours" etc. Something along those lines. That would be the ideal situation and I would not say a word about it. Next possibility, which is not ideal, but acceptable, would be if they didnt post anything. I would simply think - yes, th are probably very busy working on it, trying to fix it, and dont have time to post on FB. Fair enough. However, what I see as extremely arrogant and unacceptable is that everytime this happened, their FB page was full of posts from other people asking what is happening. And they posted nothing about this, but instead, they kept on posting some, at that point absolutely irrelevant, info about weather around the world etc., completely ignoring all people´s comments/questions and hour later, another irrelevant post from them. This IMHO shows they absolutely dont care about their users.

Another example of this is that about a yr ago, there was a major problem with the Czech translation. The problem was in the phrase "rapid fire". In English, the word "fire" has several meanings, you can fire a bullet, there can be a forestfire etc. In Czech however, the word "fire" means just the actual phenomena. They probably used Google Translate so the expression rapid fire was translated literally as "rapidly spreading fire". This expression is shown next to the name of each station which uses it. Now as you can imagine, this is a major problem because most people using WU in Czech have no idea what the original phrase was and what it really meant. So in the summer it would be very misleading, people might have thought there is a forestfire nearby. In the winter they would think WU has gone crazy, showing fires at temperatures below freezing point.
I wrote to WU via their contact form, FB, Twitter, email... I tried all channels and never got any reply whatsoever. Nothing. Only when I then posted here on the forum about it, Ken, our great forum admin, knows some guy at WU, so he asked him and lo and behold, next day it was fixed - only showing that it was not a technical problem to fix it, it was just pure ignorance...

3. data accuracy.
Being someone who loves statistics and data accuracy, some of the techniqes WU uses drive me nuts. Example - a common scenario - imagine a day where most of the day it is overcast, lets say the temperature ranges between 20 to 22°C the whole day, only for a very short while in the afternoon, the clouds break and you can see the sun for about an hour, which causes the temperature to rise to about 28°C. Lets ignore night just for simplicity. So the daily min would be something like 20, daily maximum that 28.
Now what would you say is the average temperature for that day? I say it would be something like 23, slightly above that 20-22, due to the short period of 28. On the other hand, based on WU, the average temperature will be 24, and in some cases this difference will be even much larger. Despite the fact they have all your daily values, so they could simply calculate the mean the way it is supposed to be calculated (sum of all Ts divided by the number of measurements), what they do is simply use the middle value between daily min and daily max. This is again in my opinion very misleading and inaccurate.

Offline Bashy

  • brecklandweather.com/meteo
  • Forecaster
  • *****
  • Posts: 1465
  • brecklandweather.com/index.php
    • Breckland Weather
Re: WU website down
« Reply #45 on: June 25, 2017, 08:08:47 AM »
Thanks for clarifying, it is good to know why

1) Yes we are classed as a minority and they go with the majority, i can understand that, the majority  also includes local Metars too. I have been in the situation myself where i was the minority when i had just the OS station and i was -2 below all other stations. That turned out to be 100% correct in WU's favour, my station was under reading, not had the problem since with the Davis/

I have also been in to it with Awkeas, when i had just the OS station, the blocked all my instruments because i was sending solar data, they said no other OS station is sending solar, OS does not have solar, yet i had the Solar in a jar modification, they would not listen. At that point i just walked away from them, but, i do not continually moan about them ;)

2) all i can say is, support is very slow, sometimes 2 weeks later, why, i do not know, perhaps only one person deals with support and its always been Kari that has responded to me, i have had recent communication from Dan though as i suspect many if not all have had.

3)Never noticed, does not interest me, it is of no value to me, all i care about is the Min and Max, why is the mean temp so important? I spose its more for climate usage? again, the only interest i have in that side of things is with regards to extreme records

Did you fill in their survey the other week when i left the url to it? any issues, that was the place to let them know.
Its a free service they provide, i just take it as it is, if it goes offline thats their problem, they have more to lose than i do.

Why dont you pen an email to Dan with your concerns, ask why the algorithm works like it does, also voice your concern regarding the tin can stations, i can see where you are coming from with that, but i also think they base it on the nearest metar too, if that also relates to the other stations then i can understand them choosing those over you perse' if no metar is included then that could be a problem.

But, it is nice to know what ya keep shouting about, i wonder if the others have different issues?
Kind regards
Bashy

Offline Jáchym

  • Meteotemplate Developer
  • Forecaster
  • *****
  • Posts: 8605
    • Meteotemplate
Re: WU website down
« Reply #46 on: June 25, 2017, 08:27:12 AM »
Why I dont send any email to WU?

Read my second point... I think it is pretty obvious...I dont have time to waste to be quite honest...

And also, the fact you dont care about avg temperature is fine, but that still doesnt make it any more justifiable for them to do this. If they want to show avg T, then they should calculate it the way it is supposed to be calculated and I hope no-one will argue with me that middle value between min/max is not an average.

Offline Bashy

  • brecklandweather.com/meteo
  • Forecaster
  • *****
  • Posts: 1465
  • brecklandweather.com/index.php
    • Breckland Weather
Re: WU website down
« Reply #47 on: June 25, 2017, 08:42:17 AM »
But, just moaning about them will not get it sorted, actions speak louder than words.....
Kind regards
Bashy

Offline Jáchym

  • Meteotemplate Developer
  • Forecaster
  • *****
  • Posts: 8605
    • Meteotemplate
Re: WU website down
« Reply #48 on: June 25, 2017, 08:56:25 AM »
Bashy I am not moaning, I was only explaining the reasons for why I dont like WU - because I got some emails where people asked me about this.

And I completely agree that actions speak louder than words, but there is a point where you just have to say "this is a waste of time". Just imagine you wanted some block, you email me about it and I completely ignore your email. You might do it one more time again some time later. But after that you are most likely going to give up as well because it would just be waste of time.

I was willing to help them, I tried contacting them several times, via several different communication channels, offering suggestions how the problem can be solved. Never got a single reply and when I then saw the attitude on FB, which also happened repeatedly, I just thought to myself, who cares... it is their problem.

So I only posted the above to explain the reasons why I dont like that website at all and the only reason I actually still look at it and use it at times is because of you - the Meteotemplate users, because some of you dont see things so bad and want to use it. I could have said I will not use WU at all and just ignore it, but as you know, I didnt do that and because people wanted it, I made several scripts which use it, including the update and data import.

Offline Bashy

  • brecklandweather.com/meteo
  • Forecaster
  • *****
  • Posts: 1465
  • brecklandweather.com/index.php
    • Breckland Weather
Re: WU website down
« Reply #49 on: June 25, 2017, 09:06:11 AM »
Every WU issue post I have read there has been a pack of moaners, about I how much they hate it and "you will so find how bad it is"  etc etc.  Not exact words but ya get me drift, I just don't understand why the necessity to constantly moan about it, that's all.
Kind regards
Bashy