New Weather Data Site

[Donald Giuliano]

David Demko and I would like to announce a new weather data website we are operating at http://www.twisterdata.com. We weren't sure where to post this, so we decided to place the announcement in the Marketplace. This is because we do have advertisements on the site, which we're using as a source of revenue to help cover the cost of the souped-up 8-core dedicated server and 3 TB/month bandwidth we've leased for this project. (MODS: If this post should be in another location, feel free to move.)

At the site you will find a wide array of standard model output parameters, such as Temp/Wind/CAPE/CIN/LCL/etc. Right now the only data products available are from the NAM, but we plan on adding the RUC and GFS soon, as well as the ability to loop and compare images and soundings.

Current website features include:

• Easy navigation with a user-friendly interface
• Large, easy-to-read, high-resolution graphics
• Forecast soundings (approx. 1 sounding per county in the Plains)
• All forecast graphics available in 3-hour increments
• Archive of the last 4 days of model runs
• Ability to change some default website settings (saved to a cookie)

Please note that due to the large, high-resolution nature of the graphics, having a monitor resolution of 1280x1024 and a reasonably fast internet connection makes the site a lot snappier. We hope everyone enjoys the website and finds it useful. Comments and suggestions are welcome.

For more details visit the website tutorial.

--Don

 

[Skip Talbot]

Wow, great looking weather graphics, and a very nice website design. Is this site going to stay a free service or become subscription based?

 

[Donald Giuliano]

Wow, great looking weather graphics, and a very nice website design. Is this site going to stay a free service or become subscription based?

We plan to keep the products you see right now available for free, as well as similar content for the RUC/GFS. We may add some subscription products, such as access to archived data or the ability to create custom graphics, but have no plans to do so at present. For more information, check our about page. It doesn't say much more right now, but any changes we make will be reflected there.

 

[Steve Miller TX]

Nicely done! The graphic details and color schemes are very nice and easy to read. Interface is intuitive and easy to use. Plus, we get the 6/18z runs of the NAM too! I can't wait to see your other model products you have planned.

 

[Steve Miller]

I'll second what SMTX says!
This TVN offering is nicely done. The products load fast and are prepared well, i.e. ease of viewing data and making sense of it in short order. I too look forward to seeing the other models added. Are the soundings working for anyone else?

I'm excited to see new resources like this and spotterchat.org (also announced on ST today) and look forward to seeing what else will be introduced over the winter!

 

[Stuart Robinson]

Interesting site – I like the sounding tool that you have.
If you don’t mind can I put forward some constructive criticism?
Firstly I note that you have missed out some basic search engine meta tags – while the debate over there usefulness continues. I am still a firm believer that you are better to have them than not. Try inputting your site into the following tools.

http://www.submitexpress.com/analyzer/
http://www.webconfs.com/keyword-density-checker.php

Now I Really like the sounding map – but it is actually quite a pain to use. What would be good is that if you can click on the goggle map, you can then redraw the Skew-T for that location. Hint – Google map can actually return a LAT/LON coordinate on mouse click and you could use that to relocate the Skew-T.

The plotted map images are too big and this is noticeable even over a high speed internet connection – over a EVDO card, out in the field it would be a real pain. Can you reduce the image sizes? 300kb is way too big.

 

[Tony Perkins]

This looks very nice. I have another comment to add onto Stuart's about size. Since a lot of chase laptops have a widescreen that are not very tall pixelwise, I would find myself having to scroll down to the see the bottom half of a map. A second gfx mode where the gfx are shorter would be a nice addition.

 

[Dennis Dennison]

Excellent! This site of yours is well done, and has a ton of info!
Great Job

 

[Donald Giuliano]

These are some excellent suggestions, guys, we really appreciate both your compliments and criticism. Here are some answers to questions you've had so far...

Interesting site – I like the sounding tool that you have.
If you don’t mind can I put forward some constructive criticism?
Firstly I note that you have missed out some basic search engine meta tags – while the debate over there usefulness continues. I am still a firm believer that you are better to have them than not. Try inputting your site into the following tools.

http://www.submitexpress.com/analyzer/
http://www.webconfs.com/keyword-density-checker.php

I'll be honest, when it comes to SEO we know next to nothing. I'll definitely look into these websites.

Now I Really like the sounding map – but it is actually quite a pain to use. What would be good is that if you can click on the goggle map, you can then redraw the Skew-T for that location. Hint – Google map can actually return a LAT/LON coordinate on mouse click and you could use that to relocate the Skew-T.

We're actually planning on doing this, as it would make navigation a lot easier. This feature is probably one of the first we'll tackle, since it should be relatively easy to implement.

The plotted map images are too big and this is noticeable even over a high speed internet connection – over a EVDO card, out in the field it would be a real pain. Can you reduce the image sizes? 300kb is way too big.

This looks very nice. I have another comment to add onto Stuart's about size. Since a lot of chase laptops have a widescreen that are not very tall pixelwise, I would find myself having to scroll down to the see the bottom half of a map. A second gfx mode where the gfx are shorter would be a nice addition.

This is something we have been concerned about, since we have really focused on producing the highest-quality images we can feasibly make. We worked quite a bit on optimizing the image sizes, but we can't make the file sizes any smaller than about 150-300K (depending on the map) at that resolution without substantially degrading the quality (e.g., by using JPEGs instead of PNGs). Creating an option to view the images at smaller file/resolution sizes is not that difficult, so we'll definitely consider this in the future. However, since chase season is still a ways off and people are not likely to be using EVDO cards to access data much until then, we're probably going to focus on getting more models online first (such as the RUC and GFS) before designing a "chase mode" into the website. This suggestion, though, is duly noted.

 

[Kathryn Piotrowski]

Well done! This is going on the top of my favorites list for severe weather data.

 

[Bob Hartig]

I really like what you've done. Thanks for making it available; it's something I will definitely be using in the future.

 

[Donald Giuliano]

Thanks Kathryn and Bob for your compliments, we're glad you like the site.

Now I have a question for those who have viewed the TwisterData.com website. What do you think about the terrain shading? We're kind of wondering how useful it is, and whether or not it's worth keeping. Keep in mind that eliminating the terrain shading would decrease the average file size by a substantial margin, probably around 30-50%, without reducing the resolution of the maps at all.

We have three options, really:

1. Keep things as they are.
2. Eliminate terrain shading entirely.
3. Eliminate terrain shading on some maps (e.g., 300 mb wind), but keep it for others (e.g., surface temp).

Just curious to get input from some of those outside our immediate circle that have checked out the data plots.

--Don

 

[Aaron Wesson]

Very nice site! I have it marked as a favorite as well.

There is a site that I use all of the time called http://www.txtornado.net/menu.htm. She has a javascript that will allow you to mouse over the times for any level and the image will change, thus eliminating the need to click +/-6 hrs.
here is the 18z NAM run ---> http://www.txtornado.net/NAM/_all_nam_18z.htm

What I love about this site is that
1. it is simple and easy to use
2. It allows me to see the atmosphere on one page. I can then just chose the forecast time and see all levels at once.

This would be a great addition to your site...maybe to subscribers?

Overall very nice work on such a powerful site! As to the terrain...I can live without it.

 

[Chris Kennedy]

Amazing site! what are the chances of a satellite and radar map? and how about zoom capabilities?

 

[Jason Boggs]

Thanks Kathryn and Bob for your compliments, we're glad you like the site.

Now I have a question for those who have viewed the TwisterData.com website. What do you think about the terrain shading? We're kind of wondering how useful it is, and whether or not it's worth keeping. Keep in mind that eliminating the terrain shading would decrease the average file size by a substantial margin, probably around 30-50%, without reducing the resolution of the maps at all.

We have three options, really:

1. Keep things as they are.
2. Eliminate terrain shading entirely.
3. Eliminate terrain shading on some maps (e.g., 300 mb wind), but keep it for others (e.g., surface temp).

Just curious to get input from some of those outside our immediate circle that have checked out the data plots.

--Don

I honestly don't see a need for terrain shading, especially if it will reduce the file size. Other than that, everything looks really good.

 

[Donald Giuliano]

Amazing site! what are the chances of a satellite and radar map? and how about zoom capabilities?

We're focused on getting the models working first, and will think about doing observational data later. Dynamic zoom capabilities are unfortunately not possible with static images rendered on the server side, however we are considering something along the lines of a zoomed custom WRF run centered on risk areas (much like the Stormtrack WRF Tim linked to last week). Thanks for the suggestions!

--Don

 

[Steve Miller]

Yeah, I agree. Terrain shading on any map is not necessary. It's cool, but not necessary.
Is there a way to detect a user's connection and load maps accordingly? From home maps load almost instantly. I'll be out today and I'll give them a try on my sprint card but I suspect they will be annoyingly slow.

Once everything is squared away with the rendering, how quick do you think maps will display after a new model run?

 

[Stuart Robinson]

Simple !

A tick box to show / remove terrian from plots

 

[Steve Miller]

Genius! Here I am trying to make it harder on them - a tick box would be perfect!

 

[Johnathan Brouwer]

This is a good site. I appreciate the work you put into it.

I am curious, are you owned or funded by tornadovideos.net? They seem to be promoting you as theirs and was wondering the connection and what "being a component" means...

 

[Chris Foltz]

Don,

The site looks great. Any plans to create a floating sector or two centered on an area of potentially significant weather similar to what SPC does with their mesoanalysis page? Would make the isopleths on a few of the graphics a little easier to discern. Great job with what you guys have done so far. Looking forward to it's continued expansion.

EDIT: Guess I should have read all of the posts first as I see you've already answered a similar question.

 

[Donald Giuliano]

Once everything is squared away with the rendering, how quick do you think maps will display after a new model run?

We process the maps as soon as we get the data for a particular forecast hour and make them available immediately upon completion. In the case of the NAM, I think this takes about 60-90 seconds, so it's something like a 2-3 minute lag from the time the data appears. This may not always be the case, but I've noticed that we usually have our maps completed and available slightly before the NCEP website does.

This is a good site. I appreciate the work you put into it.

I am curious, are you owned or funded by tornadovideos.net? They seem to be promoting you as theirs and was wondering the connection and what "being a component" means...

Thanks for the compliment, we're glad you like it.

About your question, I'm not sure how we could be more clear that we are coordinating with TVN on this project. At least not without getting into the nuts and bolts of private business arrangements, which is not really appropriate.

 

[Donald Giuliano]

Now I Really like the sounding map – but it is actually quite a pain to use. What would be good is that if you can click on the goggle map, you can then redraw the Skew-T for that location. Hint – Google map can actually return a LAT/LON coordinate on mouse click and you could use that to relocate the Skew-T.

We've implemented this feature, so it should be much easier to navigate while looking at soundings. Hope you like it!

 

[Stuart Robinson]

Hi Donald,

Yes that works well. I can drag the Google map pointer to a new location and then reselect the plot and all is great - well done

 

[Stuart Robinson]

Page load time / redraw is still a tad poor - I see that you are running on an Apache server - have you enabled compression? (This is a quick win)

Also are you redrawing images on the fly directly from model data? Perhaps investigation into server image caching might help.

Over all the site is looking good and more important useful! I know over time you will add more features and Iron out bugs etc but for now you need to work on page load times