Welcome to the Cumulus Support forum.

Latest Cumulus MX release v3.1.2 build 3055 - 4 November 2019.
Legacy Cumulus 1 release v1.9.4 (build 1099) - 28 November 2014

See the Wiki for downloading either version.

Cumulusutils

Other discussion about creating web sites for Cumulus that doesn't have a specific subforum

Moderator: daj

User avatar
HansR
Posts: 344
Joined: Sat 20 Oct 2012 6:53 am
Weather Station: Davis Vantage Pro 2+
Operating System: Raspbian GNU/Linux 9 (stretch)
Location: Wagenborgen (NL)
Contact:

Re: Cumulusutils

Post by HansR » Wed 23 Oct 2019 6:49 am

HansR wrote:
Tue 22 Oct 2019 8:18 pm
(Note: the graph was inspired by this one.)
In my previous answer I gave the graph which kind of inspired mine. Notice the gray lines and areas around the median and average lines. They are all shades of grey but have a meaning: the represent statistics around a mean or the median.

Now, I don't have any statistical data around the mean I find in the Cumulus ini. It is simply a line. And I don't want to create an impression that I have more info. Simply creating a grey area is meaningless. And @beteljuice: the series are actually already stacked in the order in which they are created. If you look well, you will see where lines cross that one is drawn on top of the other. So the drawing order is already from normal to the last year, in that order.

What I will do - have to find out how first - is make a select all and none option [maybe where the average and the last year remain always visible]. It is than up to the user to start comparing years with average or whatever. The graph was never intended to give meaning with a show all If you have more than four years and activate them all, the graph often becomes very confusing and does not really give information. Using a grey area and drawing order won't change that.
https://meteo-wagenborgen.nl / Cumulus MX 3.1.0-b3053
Davis Vantage Pro 2+ / Raspberry Pi 3 Model B Plus Rev 1.3
Raspbian GNU/Linux 10 (buster) / Linux 4.19.75-v7+ armv7l
Mono JIT compiler version 6.4.0.198 (tarball Tue Sep 24 01:45:47 UTC 2019)

User avatar
beteljuice
Posts: 2770
Joined: Tue 09 Dec 2008 1:37 pm
Weather Station: None !
Operating System: ex XP SP3 - now W10 laptop grr
Location: Dudley, West Midlands, UK

Re: Cumulusutils

Post by beteljuice » Wed 23 Oct 2019 10:50 am

@Hans ...
All I was saying was that if you need to overlay (because of colour or style) in a different order to coding sequence. You don't need to re-order the code, just apply a zIndex.

Not something everybody knows about ;)
Image
......................Imagine, what you will KNOW tomorrow !

User avatar
HansR
Posts: 344
Joined: Sat 20 Oct 2012 6:53 am
Weather Station: Davis Vantage Pro 2+
Operating System: Raspbian GNU/Linux 9 (stretch)
Location: Wagenborgen (NL)
Contact:

Re: Cumulusutils

Post by HansR » Wed 23 Oct 2019 11:26 am

@Beteljuice: Ah, OK. Thnx. No misunderstanding then. And btw, I found a HighCharts CheatSheet with all possible parameters and subparameters. Very illuminating and referential :) Can't have it all in your head, can you?
https://meteo-wagenborgen.nl / Cumulus MX 3.1.0-b3053
Davis Vantage Pro 2+ / Raspberry Pi 3 Model B Plus Rev 1.3
Raspbian GNU/Linux 10 (buster) / Linux 4.19.75-v7+ armv7l
Mono JIT compiler version 6.4.0.198 (tarball Tue Sep 24 01:45:47 UTC 2019)

Phil23
Posts: 296
Joined: Sat 16 Jul 2016 11:59 pm
Weather Station: Davis Vantage Pro2 Plus
Operating System: Win7 Pro
Location: Australia

Re: Cumulusutils

Post by Phil23 » Wed 23 Oct 2019 8:56 pm

HansR wrote:
Wed 23 Oct 2019 6:49 am
(Note: the graph was inspired by this one.)
Very Impressive chart, easily visually digested.
Simply creating a grey area is meaningless.
Mostly what I see with it is the immediate identification of the data vs the trend.
Takes away the viewers tendance to have the eye revisit the legend.

I don't look at it as an average with a range of "X", it simply provides the mental & visual separation.

Personal thing maybe.

Played with the idea again & have currently modified my page for a wider line.
(In an effort to be my arguementive self, Lol).

Very simple change, but personally appealing to the mind.

Code: Select all

    series: [{
       name : 'Normal', color: 'lightgray', lineWidth: 24,
Mine will reset in about an hour unless I edit again so dropped the images below.

Phil.

GraphT.JPG
GraphR.JPG
Edit:,

While I was in the editor, decided to play with the Daily Rain as well.
Not overly impressed with this result, it could use more smoothing, sizing, colour change etc, but I did feel the 1pt line was a bit small & steppy.
But again, not that impressed with the below.
GraphDR.JPG
You do not have the required permissions to view the files attached to this post.
Image

User avatar
HansR
Posts: 344
Joined: Sat 20 Oct 2012 6:53 am
Weather Station: Davis Vantage Pro 2+
Operating System: Raspbian GNU/Linux 9 (stretch)
Location: Wagenborgen (NL)
Contact:

Re: Cumulusutils

Post by HansR » Thu 24 Oct 2019 6:09 am

@Phil: Thnx for the effort. I guess you like thick lines :lol:

But to be serious: I regard the grey zones in the chart I gave as an example not just grey zones like that. It may not be standard, but it is a well respected way to represent statistics in a graph and I will not use that technique if I don't have statistics. I don't think it adds something. As I may add statistics in future and want to be consistent over all graphing, I will save that for the real statistics. For me that is end of discussion.

But I agree with you 1) the colouring of the lines is not so very good (because of the brown line which interferes too strongly with the black as you noticed correctly) and 2) if you already have many series you get a kind of tangled skein (i found a new word :!: ) if all series are presented and you have to switch them off before you can make sense of the graph.

So I did the following:

1) I changed the colouring of the series (normal remains black). In the next release you will see: "['crimson', 'darkcyan', 'cyan', 'blue', 'green', 'yellow', 'red', 'blueviolet', 'chartreuse', 'coral', 'cornflowerblue', 'darkblue', 'darkgreen']". If you wish you can start trying out in the editor and tell me what you like/dislike. I do not really have a preference but just looking for good separation of the series. I may even think about taking this as a parameter so the user can define it's own colours. That will effectively put an end to that discussion :) I may add these colours to the Daily Rain graph as well.
2) To prevent the initial entanglement, I created an ini-parameter which governs the number of series which show up at the drawing of the graph. The additional series you want to see you have to activate yourself. Default is 2, a large number shows all. Normal always shows.

The moving average in the Daily Rain graph is indeed a bit 'steppy'. Yet I defined it as a spline. Point is, that a smoothed line is drawn between points and there are so many points in this graph that the smoothing disappears and becomes useless. The 'steppyness' is a consequence of the algorithm. Can't change that I guess, maybe scaling will help. I will look into that.

[edit :] check out the image. I enlarged the scale of the moving average and thickened the line : 2pts. You loose something but gain something else. I don't think I will do this btw ;)
You do not have the required permissions to view the files attached to this post.
https://meteo-wagenborgen.nl / Cumulus MX 3.1.0-b3053
Davis Vantage Pro 2+ / Raspberry Pi 3 Model B Plus Rev 1.3
Raspbian GNU/Linux 10 (buster) / Linux 4.19.75-v7+ armv7l
Mono JIT compiler version 6.4.0.198 (tarball Tue Sep 24 01:45:47 UTC 2019)

User avatar
HansR
Posts: 344
Joined: Sat 20 Oct 2012 6:53 am
Weather Station: Davis Vantage Pro 2+
Operating System: Raspbian GNU/Linux 9 (stretch)
Location: Wagenborgen (NL)
Contact:

Re: Cumulusutils

Post by HansR » Fri 25 Oct 2019 12:59 pm

Version 1.7.1 is online
This is a maintenance version concerning the Graphs module.
Introduction of user defined colours for line graphs.
https://meteo-wagenborgen.nl / Cumulus MX 3.1.0-b3053
Davis Vantage Pro 2+ / Raspberry Pi 3 Model B Plus Rev 1.3
Raspbian GNU/Linux 10 (buster) / Linux 4.19.75-v7+ armv7l
Mono JIT compiler version 6.4.0.198 (tarball Tue Sep 24 01:45:47 UTC 2019)

Phil23
Posts: 296
Joined: Sat 16 Jul 2016 11:59 pm
Weather Station: Davis Vantage Pro2 Plus
Operating System: Win7 Pro
Location: Australia

Re: Cumulusutils

Post by Phil23 » Fri 25 Oct 2019 10:21 pm

HansR wrote:
Fri 25 Oct 2019 12:59 pm
Version 1.7.1 is online
This is a maintenance version concerning the Graphs module.

Code: Select all

MaxNrOfSeriesVisibileInGraph=2
Wondering if the behaviour of this setting can be modified so it works backwards from most recent.
By default I now see 2016 & 2017, but would be more logical to be seeing 2018 & 2019 as default when the option is set to 2.

Phil
Image

User avatar
HansR
Posts: 344
Joined: Sat 20 Oct 2012 6:53 am
Weather Station: Davis Vantage Pro 2+
Operating System: Raspbian GNU/Linux 9 (stretch)
Location: Wagenborgen (NL)
Contact:

Re: Cumulusutils

Post by HansR » Sat 26 Oct 2019 7:21 am

Phil23 wrote:
Fri 25 Oct 2019 10:21 pm
Wondering if the behaviour of this setting can be modified so it works backwards from most recent.
By default I now see 2016 & 2017, but would be more logical to be seeing 2018 & 2019 as default when the option is set to 2.
Well, that should not be too difficult.
I nobody shouts the opposite in the coming days, I'll implement it :)
https://meteo-wagenborgen.nl / Cumulus MX 3.1.0-b3053
Davis Vantage Pro 2+ / Raspberry Pi 3 Model B Plus Rev 1.3
Raspbian GNU/Linux 10 (buster) / Linux 4.19.75-v7+ armv7l
Mono JIT compiler version 6.4.0.198 (tarball Tue Sep 24 01:45:47 UTC 2019)

hvd51
Posts: 11
Joined: Fri 13 Jul 2012 5:05 pm
Weather Station: Davis VP2 +
Operating System: Windows 10 Ultimate
Location: 't Zandt, The Netherlands
Contact:

Re: Cumulusutils

Post by hvd51 » Sun 27 Oct 2019 11:00 am

HansR wrote:
Sat 26 Oct 2019 7:21 am
Phil23 wrote:
Fri 25 Oct 2019 10:21 pm
Wondering if the behaviour of this setting can be modified so it works backwards from most recent.
By default I now see 2016 & 2017, but would be more logical to be seeing 2018 & 2019 as default when the option is set to 2.
Well, that should not be too difficult.
I nobody shouts the opposite in the coming days, I'll implement it :)
I agree !

Harm

User avatar
HansR
Posts: 344
Joined: Sat 20 Oct 2012 6:53 am
Weather Station: Davis Vantage Pro 2+
Operating System: Raspbian GNU/Linux 9 (stretch)
Location: Wagenborgen (NL)
Contact:

Re: Cumulusutils

Post by HansR » Sun 27 Oct 2019 2:42 pm

@Phil:
@Harm:
OK, I'll implement it for the next release, no problem.

The other thing is a conversation I had with MeteoSangonera today and with Phil a bit longer ago, about predictions of the fire weather index. Both are looking for a figure for today and maybe tomorrow (which can be extended as long as there predictions of course ;) ). As it is now, pwsFWI seems always to be one day behind.

From the communication with MeteoSangonera:
HansR wrote:
Sun 27 Oct 2019 9:37 am
meteosangonera wrote:
Sun 27 Oct 2019 8:39 am
Could the weather prediction be used to obtain the predictable value of the current day?
Yes, you can always make a prediction and add it to the list for the current day. The question is: does it have any value?
However (I discussed this as well with Phil, who talked about an hourly prediction):
  1. I need a good predictive system which works well worldwide: pwsFWI must work the same everywhere worldwide, it does now and it must do so in prediction.
  2. I don't think fireweather indices are as fast changing as weather measurements because of an indirect influence on the wood. Drying and wetting is a slower process than the changing of the weather.
  3. I think weather predictions are often far off from local measurements, that may give erratic FWI predictions. Specifically an average wind speed and lowest humidity, in my opinion, will deviate much from the local measurement because of differences in measurement and calculation. And then I don't even talk about rain prediction (amount in mm).
meteosangonera wrote:
Sun 27 Oct 2019 8:39 am
I know a very good meteorological api, free of charge, from which the necessary variables could be extracted.
I think a predictive system for pwsFWI is extremely volatile and will not have great value. But however, since you are the second to ask for it: we could give it a try, predictions always come without guarantee ;)

Can you tell me which api you propose?
Do you mind if we take this discussion to the forum?
So here we are, I said I'd give it a try. MeteoSangonera came with the https://www.yourweather.co.uk/ to take the forecast. The API is relatively simple (example for my place) and can be decoded into the required parameters. I can add one, maybe even all seven days and continue the calculations.

As I cannot fill in all API-URLs for the specific locations, the user - you who runs CumulusUtils - would be required to take an account at yourweather.co.uk and to look up the URL for his/her location. I'll create an ini-parameter for that URL. If it is filled in, the forecast for the pwsFWI will be created and shown on the page. If the URL is empty or erroneous, no forecast will be generated.

I will also, in case of generated forecast, generate a text which tells the user that the pwsFWI value of today into the future is based on forecast blabla etc... So it has to be used with care and no decisions should be based etc... blabla... This will be in the Strings Section and you will have to translate that sentence if you want.

I think the above says it all.
If you have no comments I'll implement this. If you think of a better prediction, better way of implementation. Let me know. No guarantee for implementation, but it might generate other idea's

Best, Hans
https://meteo-wagenborgen.nl / Cumulus MX 3.1.0-b3053
Davis Vantage Pro 2+ / Raspberry Pi 3 Model B Plus Rev 1.3
Raspbian GNU/Linux 10 (buster) / Linux 4.19.75-v7+ armv7l
Mono JIT compiler version 6.4.0.198 (tarball Tue Sep 24 01:45:47 UTC 2019)

User avatar
HansR
Posts: 344
Joined: Sat 20 Oct 2012 6:53 am
Weather Station: Davis Vantage Pro 2+
Operating System: Raspbian GNU/Linux 9 (stretch)
Location: Wagenborgen (NL)
Contact:

Re: Cumulusutils

Post by HansR » Sun 27 Oct 2019 2:43 pm

And btw: I also started thinking about creating a Cumulus world map. That would mean that everybody who runs that specific module would participate in the automatic generation of a map containing all PWSs running CumulusUtils and therefore implicitly running Cumulus. I would transmit location, Station name, website URL (if any) and maybe some other non-private date to the map. What do you think about that as a project. To be honest: I have no idea about Maps. Google seems costly, OpenMaps might do it. Long term planning btw. Anybody knows something about mapping? Please contact me to find an efficient way to do this.
https://meteo-wagenborgen.nl / Cumulus MX 3.1.0-b3053
Davis Vantage Pro 2+ / Raspberry Pi 3 Model B Plus Rev 1.3
Raspbian GNU/Linux 10 (buster) / Linux 4.19.75-v7+ armv7l
Mono JIT compiler version 6.4.0.198 (tarball Tue Sep 24 01:45:47 UTC 2019)

User avatar
HansR
Posts: 344
Joined: Sat 20 Oct 2012 6:53 am
Weather Station: Davis Vantage Pro 2+
Operating System: Raspbian GNU/Linux 9 (stretch)
Location: Wagenborgen (NL)
Contact:

Re: Cumulusutils

Post by HansR » Sun 27 Oct 2019 3:50 pm

Version 1.7.2 is online
As requested: In the Monthly Temp and Monthly Rain graphs, the MaxNrOfSeriesVisibileInGraph, now is valid for the most recent years iso the first years.
https://meteo-wagenborgen.nl / Cumulus MX 3.1.0-b3053
Davis Vantage Pro 2+ / Raspberry Pi 3 Model B Plus Rev 1.3
Raspbian GNU/Linux 10 (buster) / Linux 4.19.75-v7+ armv7l
Mono JIT compiler version 6.4.0.198 (tarball Tue Sep 24 01:45:47 UTC 2019)

Phil23
Posts: 296
Joined: Sat 16 Jul 2016 11:59 pm
Weather Station: Davis Vantage Pro2 Plus
Operating System: Win7 Pro
Location: Australia

Re: Cumulusutils

Post by Phil23 » Sun 27 Oct 2019 8:37 pm

Uploaded...

One thing I'm not quite getting is the relationship between the Series & the Colours.


I can see the list of colours, but how they are matched is unclear.
Going from the start of the list I'd assume

2016=darkcyan
2017=crimson
2018=cyan
2019=blue

Code: Select all

[Graphs]
PeriodMovingAverage=180
GraphColors=['darkcyan', 'crimson', 'cyan', 'blue', 'green', 'yellow', 'red', 'blueviolet', 'chartreuse', 'coral', 'cornflowerblue', 'darkblue', 'darkgreen']
MaxNrOfSeriesVisibileInGraph=2
My chart is more like
2016=green
2017=red
2018=cyan
2019=blue

Other very minor suggestion is the location of the Go button on the left of the drop down menu.

It seems to be in reverse to the way logic would have you using controls on a page.

Would make more sense to have it at the right on the dropdown.

Cheers

Phil.
Image

User avatar
HansR
Posts: 344
Joined: Sat 20 Oct 2012 6:53 am
Weather Station: Davis Vantage Pro 2+
Operating System: Raspbian GNU/Linux 9 (stretch)
Location: Wagenborgen (NL)
Contact:

Re: Cumulusutils

Post by HansR » Sun 27 Oct 2019 8:53 pm

Phil, it's first series: first colour, second series: second colour etc...
What you call green is really darkcyan. And one of the reasons I made these colours userdefinable is that you can experiment whatever you like.
Check out: https://www.w3schools.com/colors/colors_names.asp (and the whole of the colors tutorial.
https://meteo-wagenborgen.nl / Cumulus MX 3.1.0-b3053
Davis Vantage Pro 2+ / Raspberry Pi 3 Model B Plus Rev 1.3
Raspbian GNU/Linux 10 (buster) / Linux 4.19.75-v7+ armv7l
Mono JIT compiler version 6.4.0.198 (tarball Tue Sep 24 01:45:47 UTC 2019)

Phil23
Posts: 296
Joined: Sat 16 Jul 2016 11:59 pm
Weather Station: Davis Vantage Pro2 Plus
Operating System: Win7 Pro
Location: Australia

Re: Cumulusutils

Post by Phil23 » Sun 27 Oct 2019 10:06 pm

Ok,

After some experimenting I can see it's behaving as expected.

I do notice though that my choice of colours for Monthly Rain & Monthly Temp are causing the colours for daily rain to not be ideal.

Understand that they follow the same rules.

Phil.
Image

Post Reply