Snow forecasters comparison daily

[Star_Hawk]

Hey guys, this might be a stupid idea, please don’t hesitate to tell me if it is, but I just thought I’d put this out there. In all honesty, I’m pretty nervous about even posting this but after much deliberation I’ve decided to anyway. I’m not sure if it’s being done already, as far as I can tell it isn’t, but it’s just something I do in my spare time, so I thought It’d be worthwhile putting it here, if anyone cares.
And to be clear, I know people already forecast their own predictions on here from weather models, this would be more of a thing for people to glance at in the morning or something.

Anyways, basically in my spare time for the last 2 years I’ve been making a database each season comparing major snow forecasting websites, so that I can get an idea of what they’re each saying and get an average of all of them, to avoid just getting my info from one source (and then I compare that average with what people are saying on here). This includes forecasters like snowatch, weatherzone, mountain watch, j2ski, snow-forecast ect. So I have this huge table in excel comparing their 7 day forecasts, they’re day to day forecasts and their accuracy.

So basically, to get to the point, I was just wondering weather people would like/want a website I would make that puts this info out there for everyone to see. I’m really not sure if people would want it, but I thought why not, I may as well ask. So do tell me your thoughts (pls don’t roast me if it’s already been done). Thanks.

 

[Nidecker]

Hey guys, this might be a stupid idea, please don’t hesitate to tell me if it is, but I just thought I’d put this out there. In all honesty, I’m pretty nervous about even posting this but after much deliberation I’ve decided to anyway. I’m not sure if it’s being done already, as far as I can tell it isn’t, but it’s just something I do in my spare time, so I thought It’d be worthwhile putting it here, if anyone cares.
And to be clear, I know people already forecast their own predictions on here from weather models, this would be more of a thing for people to glance at in the morning or something.

Anyways, basically in my spare time for the last 2 years I’ve been making a database each season comparing major snow forecasting websites, so that I can get an idea of what they’re each saying and get an average of all of them, to avoid just getting my info from one source (and then I compare that average with what people are saying on here). This includes forecasters like snowatch, weatherzone, mountain watch, j2ski, snow-forecast ect. So I have this huge table in excel comparing their 7 day forecasts, they’re day to day forecasts and their accuracy.

So basically, to get to the point, I was just wondering weather people would like/want a website I would make that puts this info out there for everyone to see. I’m really not sure if people would want it, but I thought why not, I may as well ask. So do tell me your thoughts (pls don’t roast me if it’s already been done). Thanks.

Yes!

 

[Untele-whippet]

Just do it !

 

[LaNeige]

You're asking a bunch of people who a) too much time thinking about skiing b) the remaining time looking at the sites you mentioned and c) spend leftover time posting about that information. It's a yes.

 

[Kletterer]

@Jellybeans had been doing an accuracy thread here in previous seasons. This being the best weather content / predictions community for Aus ski fields- why not join us and post here ?

 

[Star_Hawk]

Ok, so it seems like people are interested. 1) Is there any other sites I should include? I’ve got weatherzone, j2ski, snow forecast, mountain watch and snowatch so far. 2) so far I’m measuring accuracy day to day, accuracy of overall snow forecast for the season vs total snowfall for the season, 7 day forecasts, single day forecasts, variation between forecasters and weighted averages for the whole lot of them combined. Is there anything else I should include? And 3) I should be able to get it online in a week or 2, if that’s good.

Oh and I have to think of a name for it.

 

[Star_Hawk]

You're asking a bunch of people who a) too much time thinking about skiing b) the remaining time looking at the sites you mentioned and c) spend leftover time posting about that information. It's a yes.

Yeah I know, but I wanted to gauge interest before putting a lot of time into making the website. I had hoped people would be interested because I already have an excel table of the data.

 

[Gargamel]

Nice work @Star_Hawk

What’s the ranking?

My Guess (most accurate as number 1, least accurate number 5).

1. Snow watch
2. Weatherzone
3. Mountainwatch
4. J2ski
5. Snow-forecast

IE the order you mentioned in your original post.

 

[lewis]

Nice work @Star_Hawk

What’s the ranking?

My Guess (most accurate as number 1, least accurate number 5).

1. Snow watch
2. Weatherzone
3. Mountainwatch
4. J2ski
5. Snow-forecast

IE the order you mentioned in your original post.

1. GFS
2. ?
3. GFS
4. ?
5. GFS

 

[Star_Hawk]

Nice work @Star_Hawk

What’s the ranking?

My Guess (most accurate as number 1, least accurate number 5).

1. Snow watch
2. Weatherzone
3. Mountainwatch
4. J2ski
5. Snow-forecast

IE the order you mentioned in your original post.

Current rankings are (% of days correct since may 2nd):
1. Snowatch - 74%
2. Weatherzone - 73%
3. J2ski - 72%
4. Snow-forecast - 64%
5. Mountain watch - 62%

 

[Gargamel]

Current rankings are (% of days correct since may 2nd):
1. Snowatch - 74%
2. Weatherzone - 73%
3. J2ski - 72%
4. Snow-forecast - 64%
5. Mountain watch - 62%

Wowsers - what a surprise I didn’t think the gap was that large.

Thanks so much for sharing

 

[Jellybeans]

@Jellybeans had been doing an accuracy thread here in previous seasons. This being the best weather content / predictions community for Aus ski fields- why not join us and post here ?

Resource - Snow prediction accuracy thread

So this kinda explains it. Yeah, thats a concern...but then again its just analysis of accuracy Simple analysis. Just numbers. Nothing else. No discussion in thread. Leave any of that for bbq. Be good to do the same for models. GFS AXS EC Synoptics. 7 days. 4 days. Analysis. Same date per post...

www.ski.com.au

Has historically been Jane/WZ/EC then GFS/CMC then ACCESS/Frog/MW IIRC.

I’m working FT at Buller so I can’t really update it anymore but others are welcome to use the data or improve on it.

 

[LaNeige]

I would add Jane to your list, but she provides a proprietary service, so you might be on dicey ground publishing her data, although I'm sure she'd love to see the stats (if she's on top).

In terms of a cheap and easy was to publish your excel spreadsheet, I'd suggest using R publications of R Shiny. Both are free and you can import your SS and publish it pretty quickly.

 

[Star_Hawk]

In terms of a cheap and easy was to publish your excel spreadsheet, I'd suggest using R publications of R Shiny. Both are free and you can import your SS and publish it pretty quickly.

I was thinking that I’d make a nice website instead of just importing an excel document. The excel doc is a little.. all over the place, because it makes sense to me but maybe not others. In terms of money and time, im okay spending money on this as it’s something i really care about, and im definitely okay putting a lot of time into it.

 

[Donza]

Personally I wouldn't bother with a website.
Your analysis comes from one weather model. Which isn't the most accurate.

 

[POW Hungry]

Good on you. I like your direction and ambition.

I tend to agree with @Donza . You'll burn your own time/money for web, web design etc. for little gain.
There's enough of your audience here so why not just add the aggregate of forecasting outcomes here - have your own thread if you want.

 

[Donza]

I tend to agree with @Donza . You'll burn your own time/money for web, web design etc. for little gain.
There's enough of your audience here so why not just add the aggregate of forecasting outcomes here - have your own thread if you want.

This.
Effort for nothing.
Gotta invite the Europeans. They know a few things about forecasts.

 

[Star_Hawk]

Personally I wouldn't bother with a website.
Your analysis comes from one weather model. Which isn't the most accurate.

Two things, 1) the websites, at least from the data I’ve collected so far, produce vastly different forecasts, for instance for the next 7 days from today WZ says they’ll be 23cm, snowatch says 14cm and j2ski says 4cm. 2) it’s just a thing I’m interested in and I thought people might find it interesting, it’s not supposed to be highly detailed intricate analysis of different models, it’s more that those websites are ones that people visit often, so I thought putting it all in once place with comparisons would be helpful.

 

[Star_Hawk]

Good on you. I like your direction and ambition.

I tend to agree with @Donza . You'll burn your own time/money for web, web design etc. for little gain.
There's enough of your audience here so why not just add the aggregate of forecasting outcomes here - have your own thread if you want.

I really enjoy making websites and making things look nice, I guess I could do a thread but I do enjoy having a project to work on.

 

[Donza]

Two things, 1) the websites, at least from the data I’ve collected so far, produce vastly different forecasts, for instance for the next 7 days from today WZ says they’ll be 23cm, snowatch says 14cm and j2ski says 4cm. 2) it’s just a thing I’m interested in and I thought people might find it interesting, it’s not supposed to be highly detailed intricate analysis of different models, it’s more that those websites are ones that people visit often, so I thought putting it all in once place with comparisons would be helpful.

Then you should probably include forecasting sites with the most accurate models.
Will get a better result.
All those sites you listed derive their data from gfs. Some are purely computer driven. Some have human input at some point.

 

[Star_Hawk]

Then you should probably include forecasting sites with the most accurate models.
Will get a better result.
All those sites you listed derive their data from gfs. Some are purely computer driven. Some have human input at some point.

Yeah, I like seeing the comparisons with the human vs computer sites. What sites do you recommend I should include? I’d love to add more accurate sites.

 

[Donza]

Yeah, I like seeing the comparisons with the human vs computer sites. What sites do you recommend I should include? I’d love to add more accurate sites.

Yr.no is a good start.
BoM.

Jane Bunn.
* you may need approval.

 

[Star_Hawk]

Personally I wouldn't bother with a website.
Your analysis comes from one weather model. Which isn't the most accurate.

And in all honesty, I’m still going to be putting the data into my excel sheet regardless of wether it’s published or not. I just thought people might find it interesting, and it might help people who browse these websites to get an average.

 

[Star_Hawk]

Yr.no is a good start.
BoM.

Jane Bunn.
* you may need approval.

Was already looking at the first two. Some one mentioned Jane, not sure what the deal is with that.

 

[rocketboy]

Just needs the KISS treatment and it could easily be a viral hit.

 

[Star_Hawk]

Yes!

Just do it !

You're asking a bunch of people who a) too much time thinking about skiing b) the remaining time looking at the sites you mentioned and c) spend leftover time posting about that information. It's a yes.

Just needs the KISS treatment and it could easily be a viral hit.

In the meantime before I publish the website (in Thredbo at the moment so terrible internet and no access to my main computer), should I make a new thread with my current data? Also making an instagram for it as that is easier for some, but that’s a WIP.

 

[POW Hungry]

P.S. I 'aint got much data to back me but I'll go out on a limb and suggest the frog (Snowatch) hasn't a pinch of Pidgeon shit on any of those other sources beyond 4-5 days.

 

[POW Hungry]

In the meantime before I publish the website (in Thredbo at the moment so terrible internet and no access to my main computer), should I make a new thread with my current data? Also making an instagram for it as that is easier for some, but that’s a WIP.

Post it here.
You've got some context & traction here, so use it.

 

[LaNeige]

I was thinking that I’d make a nice website instead of just importing an excel document. The excel doc is a little.. all over the place, because it makes sense to me but maybe not others. In terms of money and time, im okay spending money on this as it’s something i really care about, and im definitely okay putting a lot of time into it.

You can make a pretty website using the R Shiny with plenty of interactivity - it's not just a dump of a spreadsheet. I suggest this as it provides a convenient and low overhead way to get a data driven website up and running. Here's an example -
https://shiny.rstudio.com/gallery/didacting-modeling.html

 

[Star_Hawk]

Post it here.
You've got some context & traction here, so use it.

Will do tomorrow.

 

[Star_Hawk]

Oh and while I get the website thing sorted, I’ll also post the numbers on an instagram account I made for it @forecasterscompared if anyones intrested, as well as here on the forums.

 

[Ramshead]

I have long thought that a website with the ladder/table/standings/whatever you call it of every major sporting league in the world would be good. Laddersoftheworld.com. Steal the idea, I'll never do it. Aggregation, done right, can be appealing. I think Star Hawk there's not enough people with a deep interest in Aus snow weather who would make it financially viable, or even close. But if you're doing it for love and want to make a nice thing, well that's a good reason to do a thing even if not a single person but you ever sees it

 

[Star_Hawk]

I have long thought that a website with the ladder/table/standings/whatever you call it of every major sporting league in the world would be good. Laddersoftheworld.com. Steal the idea, I'll never do it. Aggregation, done right, can be appealing. I think Star Hawk there's not enough people with a deep interest in Aus snow weather who would make it financially viable, or even close. But if you're doing it for love and want to make a nice thing, well that's a good reason to do a thing even if not a single person but you ever sees it

You’ve hit the nail on the head there. I’ve been doing it for years on excel anyways, but I just put it here because I thought people, especially those who check these sites each morning, might find the data interesting or useful. I know it’s not really helpful to those like POW and Donza who already analyse the models directly, it’s not meant to be. It’s just a passion project I thought was worth sharing, so here we are. Anyways, I’ll post the tomorrows data in the morning, and see how it goes.

 

[Donza]

I have long thought that a website with the ladder/table/standings/whatever you call it of every major sporting league in the world would be good. Laddersoftheworld.com. Steal the idea, I'll never do it. Aggregation, done right, can be appealing. I think Star Hawk there's not enough people with a deep interest in Aus snow weather who would make it financially viable, or even close. But if you're doing it for love and want to make a nice thing, well that's a good reason to do a thing even if not a single person but you ever sees it

Moscow 1980
Los Angeles 1984

 

[Star_Hawk]

*am working on adding BoM and yr.no
Here’s todays numbers:

 

[skichic]

I like the simple, clean look

 

[Untele-whippet]

Nice, thx

 

[Star_Hawk]

Forecasts for July 9th

 

[Star_Hawk]

Forecasts for 10th July. Bom and yr.no should be integrated next week.

 

[LaNeige]

You asked about Jane Bunn a couple of post back and I've pasted the link below. But a quick summary is that she a real meteorologist and a keen skier who does the Channel 7 weather in Melb. And also has a lot of fans on the forums.

Also, out of curiosity where do you get your actually snow measurements from?

Jane's Update | Jane's Weather

janesweather.com

 

[Star_Hawk]

Also, out of curiosity where do you get your actually snow measurements from?

I usually use a combination of Victorian and nsw resorts. Usually is Perisher in nsw because they have a snow stake, which makes it easier to see if the resort is lying about their snowfall amount, and then falls in vic and Buller. As Perisher is close to Thredbo they usually get around the same amount of snow, and falls is close to Hotham.

 

[zzaaarp]

Nice. As a suggestion - headline with a single aggregated data point, weighted by the recent accuracies. Ie one single XXcm up in lights above the charts as the weighted average most likely outcome

 

[Star_Hawk]

Nice. As a suggestion - headline with a single aggregated data point, weighted by the recent accuracies. Ie one single XXcm up in lights above the charts as the weighted average most likely outcome

I can absolutely do that. Funny enough I already have a weighted average formula for this, I just haven’t published it yet. I will do this for tomorrow.

 

[Star_Hawk]

Forecasts for the 11th of July. Weighted average has been added.

 

[Star_Hawk]

Forecasts for the 12th of July. J2ski is a tad pessimistic with this one.

 

[janesweather]

No problem at all using my forecasts, that's why I write them.
Anyone can access at https://janesweather.com/janes-update/snow-forecast
(If you want to compare individual models, or have the updates sent to you, that's where the subscription comes in)
On janesweather it shows the output from ACCESS, GFS and GEM that we source ourselves. janebunn.net shows EURO from met.no
Perisher and Hotham still look at a spot a bit too far east from the rest of the herd (and why when it's an easterly system they have huge numbers) but we have nearly (finally) solved that problem.

 

[Star_Hawk]

No problem at all using my forecasts, that's why I write them.
Anyone can access at https://janesweather.com/janes-update/snow-forecast
(If you want to compare individual models, or have the updates sent to you, that's where the subscription comes in)
On janesweather it shows the output from ACCESS, GFS and GEM that we source ourselves. janebunn.net shows EURO from met.no
Perisher and Hotham still look at a spot a bit too far east from the rest of the herd (and why when it's an easterly system they have huge numbers) but we have nearly (finally) solved that problem.

Thanks, I’ll add your forecast into my charts in the next few days once I get back to syd from Thredbo and get fast internet back, and my main computer. I’ll definitely get the subscription as well.

 

[Star_Hawk]

I've added janes forecast to the graph and it will be in tomorrows data. Do note that due to only having 1 day forecasted in my graph the accuracy numbers for the forecast will swing around wildly for a week or 2 before stabilising.

 

[Star_Hawk]

Forecasts for the 13th of july. Janes weather has been added, note that 100% accuracy is only based of 1 data point, so it will certainly change significantly in coming days.

 

[dmz]

Forecasts for the 13th of july. Janes weather has been added, note that 100% accuracy is only based of 1 data point, so it will certainly change significantly in coming days.

Unless Jane is always right