WXforum.net

Weather Station Hardware => What Weather Station Should I Buy? => Topic started by: galfert on August 10, 2018, 11:19:23 PM

Title: Analysis of Number of Weather Stations by Brand / Type
Post by: galfert on August 10, 2018, 11:19:23 PM
MODERATORS: please don't move this thread to WU section. The purpose of this thread is not to discuss WU, but rather to inform would be buyers what is most popular going by what shows up on WU. Thanks!

I took a data export of the Weather Underground weather stations for Florida. The balance of different brands (types) might be a good indicator of other locations.

I did this twice this year. June 19 and August 10.

Here are the results. I have the data in Excel and I created custom formulas to tabulate the following.
(Jump to this post for latest numbers (Oct 2018): https://www.wxforum.net/index.php?topic=34900.msg361260#msg361260)

June 19, 2018 and August 10, 2018
WS-2902      348        419
VP2              780        546 778 Updated: error in previous formula
Ambient     1195      1253
Davis       833       827
AcuRite     1368      1365
Oregon        24          23
LaCrosse       270        275
Netatmo        26           27
WeatherFlow   38           41

* If you have an Acurite station please update your station type in Weather Underground so that we can start to differentiate Acurite by model. This is a newly offered WU Station setting option. See this link for details.
https://www.wxforum.net/index.php?topic=35425.0


I could easily add to the tabulation all others. Like Rainwise or a specific model (if so listed with WU). I could create a custom list for the whole US. That would be much more data to extract and I tried a few times and Weather Underground keeps crashing and not producing the data. I might try again some other time.

I find it interesting. Looks like the Ambient Weather WS-2902 has become a hot item. I'm impressed with it. It was also interesting to see so may VP2s go offline in about 2 months. UPDATE: there was an error in previous formula, practically no change happened to VP2 (just -2) A nearby neighbor of mine just this summer took down his ancient VP2 (had been reporting since like 2006) and replaced it with a WS-2902. I haven't talked to him to know exactly why (I've never spoken to him/her...I just noticed the switch). But his unit did have some issues this past year with not reporting rain and lots of on and off Internet time (maybe computer related - was running Cumulus).

Notes: If anyone is curious here is how I did it. I copy pasted the list of weather stations from the entire state right from WU website (it is where you read about personal weather stations and which one to buy). Takes a few minutes to generate the list. Then I paste the list into Excel. Then I add some columns. One column for each type of device I'm looking to tally. Then I fill that new column with the following formula to enter in a 1 or a 0 depending if my search string matches that row. When I'm done at the bottom of the column I just add all the 1's for the entire column. Since each column represents a given search string I label it so. Here is the example formula: =IF(ISNUMBER(SEARCH("ws-2902", D4518)), 1, 0)






Title: Re: Analysis of Number of Weather Stations by Brand / Type in FL
Post by: vreihen on August 11, 2018, 08:58:11 AM
Interesting that WeatherFlow has more installations that Netatmo, since one is a commercial product and the other is barely out of start-up phase.....
Title: Re: Analysis of Number of Weather Stations by Brand / Type in FL
Post by: galfert on August 11, 2018, 09:01:56 AM
Yep it is impossible to do the entire US in one shot. My Chrome browser ends up using over 7GB of RAM. It hangs up a lot (Not Responding). Then minutes later when the whole list is there in the browser it is impossible to copy paste that much information. My computer has 12 GB of ram and it was showing 80% used at about 9.8 GB. The list is there I just can't export it.

So I'd have to do it 50 times, state by state (or more if I include PR and DC and Guam...etc). I might do that if I have the time.
Title: Re: Analysis of Number of Weather Stations by Brand / Type in FL
Post by: galfert on August 11, 2018, 09:08:53 AM
Interesting that WeatherFlow has more installations that Netatmo, since one is a commercial product and the other is barely out of start-up phase.....

I read that WU recently removed all Netatmo stations. So maybe that is why the Netatmo numbers are so low. But apparently there is a way for Netatmo users to still get their device on WU.
Title: Re: Analysis of Number of Weather Stations by Brand / Type in FL
Post by: havtrail on August 11, 2018, 11:41:29 AM
Could you check your list to see if there are any Onset HOBO units in the untallied remainder?

Rich K.
Title: Re: Analysis of Number of Weather Stations by Brand / Type in FL
Post by: ConligWX on August 11, 2018, 11:50:11 AM
Interesting that WeatherFlow has more installations that Netatmo, since one is a commercial product and the other is barely out of start-up phase.....

I read that WU recently removed all Netatmo stations. So maybe that is why the Netatmo numbers are so low. But apparently there is a way for Netatmo users to still get their device on WU.

yep.

https://www.netatmo.com/en-GB/helpcenter/weather/1/my-weather-station-does-not-appear-on-weather-underground-is-this-normal/55
Title: Re: Analysis of Number of Weather Stations by Brand / Type in FL
Post by: openvista on August 11, 2018, 12:32:41 PM
Here's a complete list of USA WU stations as of today in CSV format. Took a bit of fussing and quite a bit of time to copy it over to Excel.
Title: Re: Analysis of Number of Weather Stations by Brand / Type in FL
Post by: galfert on August 11, 2018, 12:38:46 PM
Could you check your list to see if there are any Onset HOBO units in the untallied remainder?

Rich K.

1 Onset in PA (Used Excel Method) - I guess that is you...lone soldier in PA  \:D/
5 Onset in FL (Used Excel method)
55 in the US (just used browser Find (Ctrl-F) lookup to get this number, as it counts hits)

Be careful using Ctrl-F on the web site is not as good a method as my Excel solution. For example if you search for just "onset" then something like the town of Ponsett would trigger a false count as it contains "onset" in there. Also noticed a few stations with city names Siasconset and that too would trigger a bad count.

Because of this exercise I just noticed that I could just do the Ctrl-F instead of my complicated Excel import and formulas. But then I don't get to keep the data. I'll be back. Going to try and put some numbers together using this method for the whole US.




Title: Re: Analysis of Number of Weather Stations by Brand / Type in FL
Post by: galfert on August 11, 2018, 12:46:09 PM
Here's a complete list of USA WU stations as of today in CSV format. Took a bit of fussing and quite a bit of time to copy it over to Excel.

Nice job thank you! It isn't quite Excel though. It is coma separated values. I can still use it. I just noticed though that every item in every cell ends with a funky E with a squiggly thing on top. I can ignore it and the search formula should still work.

I'll be back....working on the data...
Title: Re: Analysis of Number of Weather Stations by Brand / Type in FL
Post by: galfert on August 11, 2018, 01:54:06 PM
Okay here it is. The numbers have been crunched...

Entire US WU Stations including territories listed below.  (Well at least entire tally for these listed types). If anyone wants to know anything else in specific let me know and I'll add it to the list. Like maybe you want to know how many Bloomsky or how many WS-1001. I can't do specifics of AcuRite because they all show up as just AcuRite Pro Weather Center. I wish AcuRite specified with more detail the station type.*

* If you have an Acurite station please update your station type in Weather Underground so that we can start to differentiate Acurite by model. This is a newly offered WU Station setting option. See this link for details.
https://www.wxforum.net/index.php?topic=35425.0


Data set: Full US List - August 11, 2018

WS-2902     7599
VP2           13338
Vue             4993
Ambient     25316
Davis         17893
AcuRite      26067
Oregon          661
LaCrosse      5222
Netatmo         428
WeatherFlow   571
Onset               58

Attached is my spreadsheet. Update: Nope couldn't attach it. It was too big and it was refused even after zipping. It was 7.4MB and this site has a 2MB attachment limit.

My Excel file could probably use some tweaking. I noticed for example that I'm not counting a very few amount of Vantage Pro 2 Stations because they are listed as "VangtagePro2" without spaces and my search term is just "Vantage Pro".  So also realize that my count of VP2 is not totally accurate because I searched for "Vantage Pro" and not "Vantage Pro2" nor "Vantage Pro 2" well because sometimes there is a space and sometimes there isn't before the number. So I just counted anything "Vantage Pro" as a VP2.  And I realize that I may mistakenly count some VP1 as VP2. Doubt very many VP1 are still operational to make that much of a difference. The search formula for Vantage Pro 2 needs to be fixed with a more complex OR in the formula. Like it needs to search for VP2 as "Vantage Pro 2" OR "Vantage Pro2" OR "VantagePro2". Also realize that there is also a very few Davis stations that are not listed as Davis ...they are listed as just Vantage Pro 2....so that affects the Davis brand totals just a bit (very few of those). If that were to be improved the formula would have to search for and count for every model of Davis as a Davis (even though it may not be called Davis). Not worth the effort I think...very few to matter. My search for LaCrosse is actually just Crosse because sometimes it is listed as La Crosse and sometimes as LaCrosse without the space, so that covers both cases, so that is good. If anyone knows Excel and they want to provide a better formula for Vantage Pro 2 stations let me know. My current formula is: =IF(ISNUMBER(SEARCH("Vantage Pro", D2)), 1, 0)

Title: Re: Analysis of Number of Weather Stations by Brand / Type
Post by: dport on August 11, 2018, 02:20:00 PM
Good stuff.  This is great analysis even with the couple anomalies you mention above.  It's too bad that WeatherFlow is gaining steam.  As I've stated before those stations are WILDLY inaccurate.  Watching a neighboring weather flow station as we speak here outside of Philly and he/she is reporting double the rain that anyone else has had today. 

Title: Re: Analysis of Number of Weather Stations by Brand / Type
Post by: galfert on August 11, 2018, 02:52:36 PM
FWIW -- Davis, Vue, VP2 what differentiation?

Not sure exactly what you are asking. Those three items are stations where the following search term was matched with the station type.

Davis = Davis
Vue = Vue
VP2 = Vantage Pro

If you add up VP2 + Vue it almost equals Davis (it is actually a bit high and I mentioned that some (very few) Vantage Pro 2 stations omit the Davis name). So that means given the formula I'm using it is pretty close to real numbers. There are a few Davis Weather Monitor stations that probably make a small bit of difference and increase the number of Davis brand number I'm reporting, and that is good...we want to count those too as Davis. It isn't perfect (again I'm open to improvements to my formula). But close enough to get a good idea of what is out there.

Title: Re: Analysis of Number of Weather Stations by Brand / Type
Post by: alanb on August 11, 2018, 03:21:18 PM
@galfert ... Very interesting information. Thanks for your efforts!
Title: Re: Analysis of Number of Weather Stations by Brand / Type
Post by: galfert on August 11, 2018, 03:43:23 PM
@galfert ... Very interesting information. Thanks for your efforts!

You are welcome. Glad you found it interesting.

In case anyone is interested in different countries I have bad news. Even though WU makes it look like they will list other countries it doesn't produce any results. Not even for our neighbors in Canada. I kinda wanted to to see what brands are most popular in other countries.

Title: Re: Analysis of Number of Weather Stations by Brand / Type in FL
Post by: havtrail on August 11, 2018, 08:21:20 PM
Could you check your list to see if there are any Onset HOBO units in the untallied remainder?

Rich K.

1 Onset in PA (Used Excel Method) - I guess that is you...lone soldier in PA  \:D/
5 Onset in FL (Used Excel method)
55 in the US (just used browser Find (Ctrl-F) lookup to get this number, as it counts hits)

Be careful using Ctrl-F on the web site is not as good a method as my Excel solution. For example if you search for just "onset" then something like the town of Ponsett would trigger a false count as it contains "onset" in there. Also noticed a few stations with city names Siasconset and that too would trigger a bad count.

Because of this exercise I just noticed that I could just do the Ctrl-F instead of my complicated Excel import and formulas. But then I don't get to keep the data. I'll be back. Going to try and put some numbers together using this method for the whole US.

Thanks so much! This is all great info and I'm sure it represents a lot of work on your part. How does it feel to be a "big data" guy? Look out, Watson...

I discovered that Onset Computer is not just a trendy, catchy name, as in the onset of something. There is a town named Onset in Massachusetts, where they are located.

Rich K.
(apparently Mr. HOBO in PA)
Title: Re: Analysis of Number of Weather Stations by Brand / Type
Post by: Mattk on August 12, 2018, 12:19:36 AM
So is WU any real indication of what is being used across all platforms and networks?
Title: Re: Analysis of Number of Weather Stations by Brand / Type
Post by: ocala on August 12, 2018, 07:42:33 AM
So is WU any real indication of what is being used across all platforms and networks?
I would certainly think so. They were for the most part the first ones to publish PWS data on a global scale. They have lost some due to their recent issues but I'm sure they represent the PWS market pretty well.
Title: Re: Analysis of Number of Weather Stations by Brand / Type
Post by: havtrail on August 12, 2018, 08:04:20 AM
So is WU any real indication of what is being used across all platforms and networks?
I would certainly think so. They were for the most part the first ones to publish PWS data on a global scale. They have lost some due to their recent issues but I'm sure they represent the PWS market pretty well.

I think there is a difference between "what is being used" and "PWS." PWS is what is says, "personal," but there are other users. The Onset HOBO products are a case in point, geared more toward research, academic and agribusiness users. Their products are a little too pricey for backyard personal weather stations.  I joined this forum because of issues with Weather Underground, but in the process have discovered that I could find no online discussion area anywhere being used by these HOBO users. It doesn't mean they are not out there. And the WXForum group is a subset that is far ahead of the typical backyard weather station person that populates WU. I don't think the general lot of PWS owners on WU are as interested, careful in siting, or as diligent as the folks on this forum. At times I wonder how many of them actually oriented their wind vanes to true north, or have their rain gauges right up against their house wall. Under these circumstances, doing "QC" by comparing everyone to these numerous stations is not a good method.

Rich K.
Title: Re: Analysis of Number of Weather Stations by Brand / Type
Post by: dport on August 12, 2018, 08:35:30 AM
Rich,

I just realized you are the person running the Onset just a couple miles from me.  I was surprised when I saw it pop up and it was an Onset HOBO.  This will be great for me to compare my station.

I’m running a VP2 w/24 hour FARS in Bryn Mawr.

https://www.wunderground.com/personal-weather-station/dashboard?ID=KPABRYNM2#history

Best,
Dan
Title: Re: Analysis of Number of Weather Stations by Brand / Type
Post by: Glenn on August 13, 2018, 12:51:35 PM
Interesting thread! Thanks for all the number crunching. It's really cool to see the actual numbers that are out there.
Title: Re: Analysis of Number of Weather Stations by Brand / Type
Post by: havtrail on August 15, 2018, 11:02:21 AM
Rich,

I just realized you are the person running the Onset just a couple miles from me.  I was surprised when I saw it pop up and it was an Onset HOBO.  This will be great for me to compare my station.

I’m running a VP2 w/24 hour FARS in Bryn Mawr.

https://www.wunderground.com/personal-weather-station/dashboard?ID=KPABRYNM2#history

Best,
Dan

Hi, neighbor! We'll have to get together sometime and compare notes. I'd like to know more about the Davis products. Our HOBO setup actually uses an adapted version of the Davis wind speed/direction sensor that Onset now offers.

If you compare our stats, I wouldn't put too much faith in my barometric pressure. I think it reports low, and I have started a comparative study against 13 closest stations on WU that surround mine.

I also have been thinking that my rain figures seemed pretty high compared to others nearby. Then in Monday's deluge I registered 5.68" for the day and Action News at 11pm reported five local suburban rainfall totals, with nearby Wayne reported at 5.58", so I feel better.

Rich K.
Title: Re: Analysis of Number of Weather Stations by Brand / Type
Post by: galfert on August 16, 2018, 09:59:01 AM
I have a favor to ask...

Can someone with a Weather Underground account from some other country (You should be browsing from that country with an IP address from there) that is not the United States look to see if they can produce a list of weather stations in your country? I was curious if the reason I can't see the list in other countries is because I'm not in that given country.

I'm hoping to hear from someone in Canada eh, or some country in Europe, or even Australia mate (only because there is probably more data, but anywhere else is great too). If you can do this, I want the data in a CSV format or Excel. You just copy it from the browser and paste it into Excel. Be patient through these steps. Let you computer do its thing even if it says Not Responding.

Here is how you get to the list:
1. Bottom of Wunderground page click on Personal Weather Stations
2. Click on "Step 2: Buy - Buying Guide"
3. Click on "More PWS Options - See More"
4. Select your country from the drop down list and click "View"
5. Wait for list to be generated
6. Copy the data with you mouse and paste into Excel, save, compress to .zip file, and post it here.

...or maybe instead of the above steps this direct link may work, then just click on Select Country drop down and copy/paste:
https://www.wunderground.com/weatherstation/hardwareandsoftware.asp

Thanks!


Title: Re: Analysis of Number of Weather Stations by Brand / Type
Post by: DW7240 on August 16, 2018, 10:39:22 AM
Here U go...…

From the U.K...


Weather Stations in United Kingdom
Station ID
Neighborhood
City
Station Type
Site
ILONDON38
Up at The O2
London
Davis Vantage Vue (Wireless)

IROSS-SH1
Maryburgh
Maryburgh, Dingwall
WMR200 Oregon Scientific


From Canada......My Location...

Weather Stations in Canada
Station ID
Neighborhood
City
Station Type
Site


YES....it's blank, guess there are non on Canada. LOL., and I find it hard to beleive that there is only two in the U.K.

Nick. dw7240.com



Title: Re: Analysis of Number of Weather Stations by Brand / Type
Post by: galfert on August 16, 2018, 11:47:44 AM
Thank you Nick.

I wonder why Weather Underground has that drop down list to search other countries and it produces no results except for the US? Weirdly enough I too get those same two stations for the UK. I guess maybe another bug in the long list of WU issues.
Title: Re: Analysis of Number of Weather Stations by Brand / Type
Post by: galfert on October 14, 2018, 02:45:18 PM
October 2018 Update:

Entire US Station List
 [ You are not allowed to view attachments ]


Just Florida
 [ You are not allowed to view attachments ]

It is particularly interesting to see the impact the recent release of the Acurite Atlas has done to these numbers.

* The Davis Vantage Pro 2 search (for the US List) was updated to include more stations picked up by the search. I'm still tweaking the VP2 formula and will update the FL list later. I'm also going to be working on updating the Davis search formula. For now at least there is some consistency with these number across all dates. When I update the formula I will recalculate previous dates. There was also a formula error in the first post of this thread. This has been corrected. Essentially there was practically no change in number of VP2 for Aug Florida totals (just -2)
Title: Re: Analysis of Number of Weather Stations by Brand / Type
Post by: galfert on October 15, 2018, 10:49:25 AM
*** New Formula for Calculating Multiple Search Matches ***

Well I finally figured it out. The problem before was that I was only matching one search result to give a tally. For example to search for all the Davis Vantage Pro 2, I was only previously looking for a match for "Vantage Pro". For Vantage Vue I was only searching for "vue". This is because some stations left out the word Davis, and others called it Vantage Pro2 or Vantage Pro II. The problem I noticed with Davis (and later with other station types too), is that there is the most common choices from the drop down selection for station type, but apparently perhaps at one point the drop down selection was not as good or even still Weather Underground may be permitting free expression of the model hardware. This lead to a few stations listed with some different names and even typos. There are stations listed as Advantage Pro. But my my search was counting those before because "vantage pro" is still a search string match for those situations. But there were also some instances like, vantagepro, davis 6153, Davis Pro 2, VP2. The typos were things like, provantage, vantage2, vantege...etc. I tried to find the most popular variations and that is why I stuck with "vantage pro" before. But now I have a formula that accounts for multiple matches. This also helped Acurite, as sometimes it was spelled Acu-rite with hyphen. Oregon Scientific also sometimes were listed as just OS WMR or WMR-xxx. Ambient was sometimes spelled Ambiant. You get the idea.

Here is the new tally using the old data with these new formulas. There is nothing terribly wrong with the previous numbers as they paint a pretty good picture of the most common name for each station type. With this new formula now we just have a better more accurate picture. Now you can almost see that if you add the VP2 + Vue stations you come pretty close to the Davis total (the balance is probably old Weather Monitors still running). The Davis total is also counting stations that are only listed as just VP2 (leaving out the word Davis) for example where as before I was only looking for the word "Davis" explicitly. This helped the Davis total go way up from the prior search method.

If anyone is curious this is the new formula (totally different than previous) for searching for multiple matches (took a while to figure this one out):
=IF(SUMPRODUCT(--(NOT(ISERR(SEARCH({"vue","view","davis 6250"},D2)))))>0,1,0)
That is an example of the Vue search. The others are formatted the same, just the search matches differ and the VP2 search is huge with so many variations.

I also noticed that the Oregon Scientific was off because it was matching positive for the State of Oregon's harware called Oregon Dept. of Environmental Quality. I had to tally those and subtract them from the Oregon total (because I was trying to pull all the Oregon models by just matching the word "Oregon".) So the Oregon Scientific total number you see reflects this correction already. You just see the Oregon Dept column because I had to tally it in order to have a number to subtract (that was easier than coming up with a custom Oregon formula with AND IF functions.)

Entire US Station List:
 [ You are not allowed to view attachments ]


Just Florida:
 [ You are not allowed to view attachments ]

This is still not perfect. When you have over 91,000 records in Excel (and multiple data sets) and you are looking and trying to account for all the variations that people mistakenly enter it gets very tedious very fast. I think these new results are good enough numbers even though probably still not perfect because of the human factor. I'm not going to sweat it trying to find every single obscure variation. I spent a lot of time on this and hopefully it was worth it to some people. Id say these number are probably 99% accurate based on the data provided by WU.

Title: Re: Analysis of Number of Weather Stations by Brand / Type
Post by: havtrail on October 15, 2018, 12:00:13 PM
Great stuff!

So, if the brand names (as opposed to individual models) now closely represent the brand totals, then in the U.S. the leader seems to be Ambient (27,384 in Oct.) followed closely by Acurite (26,762) and then Davis (19,418). Right?

This is probably the best publicly available info on the national PWS market, which is at least 91,000+ on Weather Underground alone.

Rich K.
Title: Re: Analysis of Number of Weather Stations by Brand / Type
Post by: Felix1 on October 15, 2018, 12:17:46 PM
Sorting on your raw data, looks like only 39 WeatherHawk stations are uploading their weather info to WU. I would have assumed many more than that but I guess it's because most sales are to federal, regional & local governments, universities, industry, emergency management entities and high-end residential automation.
Title: Re: Analysis of Number of Weather Stations by Brand / Type
Post by: galfert on October 15, 2018, 12:20:51 PM
Great stuff!

So, if the brand names (as opposed to individual models) now closely represent the brand totals, then in the U.S. the leader seems to be Ambient (27,384 in Oct.) followed closely by Acurite (26,762) and then Davis (19,418). Right?

This is probably the best publicly available info on the national PWS market, which is at least 91,000+ on Weather Underground alone.

Rich K.

Yes, correct. The reason the brand doesn't exactly match the listed models mostly because there are other models not listed separately and because of obscure rare variations in station types with errors or omissions.

Too bad there isn't any data for other countries. I really would have liked to have seen analysed other countries. Also too bad that Acurite and WU don't list out model numbers separately.* Seems only Davis and Ambient do this. Well I've seen the data and other manufactures also list out model numbers separately like Rainwise. I haven't bothered to include those nor others like Vaisala and Peet Bros..etc ....because nobody has asked. By looking at the numbers you can tell that +79,000 out of +91,000 is made up of just Ambient, Acurite, Davis, and LaCrosse (to a small part). Those four make up 87% of what is on WU (in the US).

* If you have an Acurite station please update your station type in Weather Underground so that we can start to differentiate Acurite by model. This is a newly offered WU Station setting option. See this link for details.
https://www.wxforum.net/index.php?topic=35425.0

PS - I also fixed Onset HOBO to include search for both instances of "onset" and "hobo."
Title: Re: Analysis of Number of Weather Stations by Brand / Type
Post by: galfert on October 15, 2018, 12:43:48 PM
Sorting on your raw data, looks like only 39 WeatherHawk stations are uploading their weather info to WU. I would have assumed many more than that but I guess it's because most sales are to federal, regional & local governments, universities, industry, emergency management entities and high-end residential automation.

I counted 54 WeatherHawk in both the August and the October data sets. Yes this analysis needs to be taken for what it is. This is what reports to WU and therefore is mostly the home owner and weather hobbyists, and there are some schools sprinkled in there too. But I think that is representative enough for the people browsing this forum and looking at what to buy for themselves. I don't think federal, local and regional governments, industry ..etc care what is on WU. This data is for us the hobbyists. But sometimes it is surprising to see some state run hardware that does show up on WU. I mentioned the Oregon Dept. of Environmental Quality, but there are others like New Jersey Dept of Environmental Protection, New Hampshire Dept. of Environmental Services...etc. But when those show up and on WU there is no way of knowing what hardware they really are using as it just shows up with those department names, because that is how they are reporting.





Title: Re: Analysis of Number of Weather Stations by Brand / Type
Post by: havtrail on October 15, 2018, 04:50:10 PM
Great stuff!

So, if the brand names (as opposed to individual models) now closely represent the brand totals, then in the U.S. the leader seems to be Ambient (27,384 in Oct.) followed closely by Acurite (26,762) and then Davis (19,418). Right?

This is probably the best publicly available info on the national PWS market, which is at least 91,000+ on Weather Underground alone.

Rich K.

Yes, correct. The reason the brand doesn't exactly match the listed models mostly because there are other models not listed separately and because of obscure rare variations in station types with errors or omissions.

Too bad there isn't any data for other countries. I really would have liked to have seen analysed other countries. Also too bad that Acurite and WU don't list out model numbers separately. Seems only Davis and Ambient do this. Well I've seen the data and other manufactures also list out model numbers separately like Rainwise. I haven't bothered to include those nor others like Vaisala and Peet Bros..etc ....because nobody has asked. By looking at the numbers you can tell that +79,000 out of +91,000 is made up of just Ambient, Acurite, Davis, and LaCrosse (to a small part). Those four make up 87% of what is on WU (in the US).

PS - I also fixed Onset HOBO to include search for both instances of "onset" and "hobo."

Back in August when the original national csv file was posted thanks to "openvista," I downloaded it and sifted through it, finding a total of 66 Onset HOBOs, reporting from 23 states. Some are just listed as "Other."

I don't think the list's choices of models is defined by the manufacturers. I think it was, as you also guessed, originally a free-form entry by the individual registrants. Later, I believe, WU instituted the drop-down choice menu with options that they created themselves, to give some level of consistency to the data. I think they included what they believed were more popular specific models in their list of choices.

I had to register our HOBO RX3003 as a HOBO U30 (an obsolete model), because that was (and still is) the only HOBO choice available. At that time I was tempted to choose "Other" instead, but went with "HOBO U30" because at least it signaled "HOBO".

Rich K.
Title: Re: Analysis of Number of Weather Stations by Brand / Type
Post by: galfert on October 15, 2018, 11:33:58 PM
Back in August when the original national csv file was posted thanks to "openvista," I downloaded it and sifted through it, finding a total of 66 Onset HOBOs, reporting from 23 states. Some are just listed as "Other."

How do you know they are Onset HOBOs if they are listed as "other"?

Quote

I don't think the list's choices of models is defined by the manufacturers. I think it was, as you also guessed, originally a free-form entry by the individual registrants. Later, I believe, WU instituted the drop-down choice menu with options that they created themselves, to give some level of consistency to the data. I think they included what they believed were more popular specific models in their list of choices.

I think some manufacturers must have some part in it though Take the WS-2000 that wasn't even officially released yet till late August and it was already a choice for model type with all of 12 stations. The WS-2000 did not fully go on sale till late August after the August WU list was pulled. This model was originally suspended for sale because a bug was discovered early on that if the SD card memory reached a certain level the device would get bricked. As a new firmware was worked on the sale was suspended. I think Ambient had to of informed WU early on that this model was new and for them to add it to the list. I think WU should clean up the list of bad choices, fix the typos and consolidate.

Quote

I had to register our HOBO RX3003 as a HOBO U30 (an obsolete model), because that was (and still is) the only HOBO choice available. At that time I was tempted to choose "Other" instead, but went with "HOBO U30" because at least it signaled "HOBO".
I would reach out to WU and see if they can add the Onset HOBO RX3003 model to the list.
Title: Re: Analysis of Number of Weather Stations by Brand / Type
Post by: galfert on November 01, 2018, 09:55:13 AM
* If you have an Acurite station please update your station type in Weather Underground so that we can start to differentiate Acurite by model. This is a newly offered WU Station setting option. See this link for details.
https://www.wxforum.net/index.php?topic=35425.0
Title: Re: Analysis of Number of Weather Stations by Brand / Type
Post by: kbellis on November 16, 2018, 01:10:13 PM
@galfert - Nicely done!

Title: Re: Analysis of Number of Weather Stations by Brand / Type
Post by: BKS97 on November 16, 2018, 02:26:08 PM
I would like to see a tally of the number of RainWise MK-111 weather stations reporting to WU.
Title: Re: Analysis of Number of Weather Stations by Brand / Type
Post by: galfert on November 16, 2018, 07:21:43 PM
It would be easy to just report how many Rainwise stations are live. Roughly about 1316 stations in the USA.  But it will take some work to first identify which of the following are MK III stations versus which are a different model. Then after we have them identified I would have to create a formula to account for each valid variation. So from the following list which are MK III and which aren't?  For example I think Rainwise CC-3000 is a an MK III but it is not coming right out and saying it is a MK III and instead it is reporting the add-on device that is doing the Internet reporting (CC-3000). But am I right or could it be a MK II or something older?

So which of these are MK III and which aren't? I don't know how to identify what should count as an MK III and which should just be a Rainwise because it might be a different model. This is more difficult than dealing with all the Davis VP2 variations only because I don't know anything about Rainwise. Correct this table and then I can look into computing some results.

Rain Wise MKIII                YES
Rain Wise MK-III-LR-RT1    YES
RainWise AgroMET              ?
RainWise CC3000                ?
Rainwise CC-3000                ?
Rainwise iP-100                   ?
Rainwise LR                   (???? is this an MK III?  or should is just be counted as a Rainwise?)
RainWise Mark III LR      YES
Rainwise Mark III              YES
RainWise Mark-III RTI-LR     YES
Rainwise MK 111-LR            YES (funny they used 1's, how dumb is that?)
RainWise MK III - LR             YES
Rainwise MK III LR               YES
Rainwise Mk III RTI LR          YES
RainWise MK III                     YES
RAINWISE MK III                     YES
Rainwise MK III-LR                     YES
Rainwise MK III-LR-RTI              YES
Rainwise MK IIIp                           YES
Rainwise MK3                                   YES
Rainwise MK-III IP100                         YES
Rainwise MKIII Long Range                  YES
Rainwise MKIII LR                             YES
Rainwise MkIII LR                         YES
Rainwise MK-III LR                           YES
Rainwise MK-III Pro                         YES
RainWise MK-III RTI                         YES
Rainwise MK-III RTI-LR                     YES
Rainwise MK-III SP                            YES
Rainwise MK-III SP1 LR                    YES
Rainwise MK-III SP1                        YES
RainWise MK-III SP-5 LR                  YES
RainWise MK-III SP-5                       YES
Rainwise MKIII w/ IP-100                 YES
Rainwise MKIII                                 YES
Rainwise MKIII-LR                           YES
Rainwise MK-IIILR                            YES
RainWise MK-III-LR                         YES
RainWise MK-III-LR-RTI                   YES
Rainwise MKIII-RTI-LR                    YES
RainWise MK-III-RTI-LR                   YES
RainWise MK-III-RTN                      YES
Rainwise MK-III-RTN-LR IP-100        YES
RainWise MK-III-RTN-LR                   YES
Rainwise MKII-LR SP1                        NO this is MK II
Rainwise MKII-LR                              NO this is MK II
Rainwise MK-lll                           YES
Rainwise MKlllLR                        YES
Rainwise MRK-III                       YES
RainWise P-100                          ?
Rainwise PVMet 300                    ?
Rainwise WS2000                        NO
Rainwise WS3000                       NO
RainWise                             (yep that is a category...do we count it as MK III model or just as Rainwise brand)
RainWise.net/ IP-100                ?
Rainwise-MKIIIRTI-LR            YES

This is a list of some of the different variations of Rainwise I've seen on WU.  Some of these in the list appear only once....some appear more frequently.
Title: Re: Analysis of Number of Weather Stations by Brand / Type
Post by: havtrail on November 17, 2018, 12:39:17 AM
You can see why they apparently have gone to a pull-down selection menu instead of freeform entry.

Rich K.
Title: Re: Analysis of Number of Weather Stations by Brand / Type
Post by: Glenn on November 18, 2018, 08:27:07 AM
Seriously nice work! I look forward to seeing the further breakdown of the Acurite stations.
Title: Re: Analysis of Number of Weather Stations by Brand / Type
Post by: galfert on April 02, 2019, 10:46:28 AM
It had been a while since I last pulled a report of weather stations on WU. I figured it was time to do it again and see how things have progressed. I was not able to pull the list.

Even though you can get this far:
https://www.wunderground.com/weatherstation/hardwareandsoftware.asp

When you try and use this:
 [ You are not allowed to view attachments ]

You end up with this:
 [ You are not allowed to view attachments ]

I hope this is just temporary and that they bring this back. This was a helpful feature to have especially as new station models came out you could gauge their popularity. You could also compare growth of brands. Although it only worked for for US stations it would have also been nice to have had access to other countries. WU please bring this back.

Title: Re: Analysis of Number of Weather Stations by Brand / Type
Post by: havtrail on April 02, 2019, 02:10:40 PM
Yes, this is a great resource, and I really appreciated the work you and others did some months back to pull down and analyze this data. It's the only way I have to track down other users of Onset weather stations and HOBOlink users. It would be nice if it worked for PWS's outside the U.S., too. I hope it comes back on line.

Rich K.
Title: Re: Analysis of Number of Weather Stations by Brand / Type
Post by: WX9TRS on May 04, 2019, 05:28:42 PM
No Peet Brothers shown. Why?  WU no longer lists them when I know there are several on the system. I do not trust there numbers. Mine is a Peet Bros 2100 but at times it says it's some other brand.  Lacrosse or something. Just plan incorrect!
Title: Re: Analysis of Number of Weather Stations by Brand / Type
Post by: galfert on May 04, 2019, 06:20:05 PM
No Peet Brothers shown. Why?  WU no longer lists them when I know there are several on the system. I do not trust there numbers. Mine is a Peet Bros 2100 but at times it says it's some other brand.  Lacrosse or something. Just plan incorrect!

No Peet Brothers because nobody ever asked. I offered to calculate any brand/model. You now asked so here it is:
All Peet Brothers branded stations or Ultimeter 800, 2000, and 2100... Total = 101 dalmatians ...I mean 101 stations (* Using October 2018 data the only thing available)

I don't see why you say you don't trust the numbers. The numbers aren't coming from Weather Underground. The numbers come from me crunching the data that Weather Underground had for station type that was something that the station owner enters. If someone enters in the wrong name or they enter in "Other" then I can't count it.

There are two different things shown when you look at a WU station. There is the Hardware and there is the Software. The Hardware never changes and like I said that depends on the station owner to enter in whatever they want. The Software on the other hand is detected and changes automatically depending on how WU sees your data being uploaded. That is part of the upload protocol that your station uses. So it is essentially up to your software to do its part there. I'm not using the software identifier. I'm only using the Hardware type entered in by the station owner. But getting back to the software the only time I've seen it change is if the station owner has inadvertently configured their upload to happen with two different software so that there is a double upload. For example with Davis you can upload from weatherlink.com or you can upload with your own 3rd party software or other logger. You want to be sure that you aren't uploading via two methods. Regardless this issue with the software is inconsequential to the process that I go through to identify the station.

Names that I matched (* Using October 2018 data..the only thing I have since this is no longer available):
PEET 2000
Peet 2100
Peet Borthers - Ultimeter 2100
Peet Bro
Peet Bros 2000
Peet Bros 2100
Peet Bros 2100
Peet Bros 2100
Peet Bros U2100
Peet Bros UI-2000
Peet Bros ULT2000
Peet Bros Ultimeter 2000
Peet Bros Ultimeter 2000
peet Bros Ultimeter 2000
Peet Bros Ultimeter 2100
Peet Bros Ultimeter 2100
Peet Bros Ultimeter 2100
Peet Bros Ultimeter 2100
Peet Bros Ultimeter 2100
Peet Bros Ultimeter 2100
Peet Bros Ultimeter 2100
Peet Bros Ultimeter 2100
Peet Bros Ultimeter 2100
Peet Bros Ultimeter II
Peet Bros Ultimeter
Peet Bros
Peet Bros
Peet Bros
Peet Bros
Peet Bros
Peet Bros. 2100
Peet Bros. Ultimeter 2000
Peet Bros. ULTIMETER 2100 Weather Station 2100
Peet Bros. Ultimeter 2100
Peet Bros. Ultimeter 2100
Peet Bros. Ultimeter 2100
Peet Bros. Ultimeter 2100
Peet Bros. Ultimeter 2100
Peet Bros. Ultimeter
Peet Bros. Utlimeter 2100 Unimount Sensor Array
Peet Bros. Weather Station
Peet Bros.
Peet Bros.
Peet Brother 2100
Peet Brotheres Ultimeter #2100
Peet Brothers 800
Peet Brothers Ultimeter 2100
Peet Brothers Ultimeter 2100
Peet Brothers Ultimeter 2100
Peet Brothers Ultimeter 2100
Peet Brothers Ultimeter 2100
Peet Brothers
Peet U2000
Peet Ultimeter 2000
Peet Ultimeter 2000
PEET Ultimeter 2000
Peet Ultimeter 2100 - Heated Sensors
Peet Ultimeter 2100
Peet Ultimeter 2100
Peet Ultimeter 2100
Peet Ultimeter 2100
Peet Ultimeter 2100
Peet Ultimeter
Peet ultimeter2100

Ultimeter 2000
Ultimeter 2000
Ultimeter 2000
Ultimeter 2000
Ultimeter 2000
Ultimeter 2000
Ultimeter 2000
Ultimeter 2000
Ultimeter 2000
Ultimeter 2000
Ultimeter 2000
Ultimeter 2000/ VWS
Ultimeter 2100 Peet Bros
Ultimeter 2100 Pro
Ultimeter 2100
Ultimeter 2100
Ultimeter 2100
Ultimeter 2100
Ultimeter 2100
Ultimeter 2100
Ultimeter 2100
Ultimeter 2100
Ultimeter 2100
Ultimeter 2100
Ultimeter 2100
Ultimeter 2100
Ultimeter 2100
Ultimeter 2100
Ultimeter 2100
Ultimeter 2100
Ultimeter 2100
ULTIMETER 2100
Ultimeter 2100
Ultimeter 2100
Ultimeter 2100
Ultimeter 2100
Ultimeter 2100-complete
Ultimeter 800
Ultimeter2100