Image with statistical information

[DelandeC]

Hello Egon,

Thank you for this new tool.

Maybe could you set another kind of green and yellow for the curves. I think this is very simillar.
What do you think about adding horizontal lines for thresholds as you reported at https://forum.blitzortung.org/showthread...6#pid21996

Thanks.

[Egon]

So that the information does not always have to be searched in the forum, I have summarized the options for parameterization on the following page. I will also record all future extensions there.

https://map.blitzortung.org/mapexample.html

/Egon

[cutty]

arrrrgghh
Cannot get "Days=" parameter to change on statistic 2... actually kills image. other parameters seem ok.

Sure wish some way to change the "graph" background color on statistic 1.  Transparency works, line color change works, but still the same graph background grey...
Mike

[Egon]

sorry, problem solved :-)

see https://map.blitzortung.org/mapexample.html for an example.

/Egon

[cutty]

sorry, problem solved :-)

see https://map.blitzortung.org/mapexample.html for an example.

/Egon

Excellent!   ....
Thanks! ... I couldn't determine what I was doing wrong!

As mentioned, I can get athe Image background change on the Statistic 1  ... it's the plot body (grey) background I'd like to modify... the graph itself.  .

AS

[Egon]

The background in the graphics area is determined by the controller status. Gray means that the station is running. Black means that there is no data. These are the 11 background colors for error messages.

   

/Egon

[cutty]

The background in the graphics area is determined by the controller status. Gray means that the station is running. Black means that there is no data. These are the 11 background colors for error messages.



/Egon

Oh for goodness sake! 
NOW I understand.... I simply didn't make the obvious connection....  sorry for the bother!!

Mike

[cutty]

The background in the graphics area is determined by the controller status. Gray means that the station is running. Black means that there is no data. These are the 11 background colors for error messages.



/Egon

Oh for goodness sake! 
NOW I understand.... I simply didn't make the obvious connection....  sorry for the bother!!

Mike

AaAH... back again
signal.php... confusion on System Blue... no issue for System Red:
Same color (yellow) set for channels 2 and 3 (a3, b1) or 3 and 4 below:
line_colors 1R,1G,1B,2R,2G,2B,3R,3G,3B,4R,4G,4B,5R,5G,5B,6R,6G,6B (default: 255,0,0,0,255,0,255,255,0,255,255,0,0,255,255,255,128,0)

Also... can't change the line colors
haven't tried any other parameters on signal.php

[Eric.Wouters]

If you set the background to transparent, there is a little issue with the header text (part that is white disappears)
add an extra modifier to set that to black or other colours ??

[cutty]

If you set the background to transparent, there is a little issue with the header text (part that is white disappears)
add an extra modifier to set that to black or other colours ??

Eric, the 'transparemt will show the background color of the element (body, cell, div, in which the lies.  If the background is white, then the white of the text will 'disappear'. 

Egon fixed the channel trace color for Ac and Ba ((2,3) on signal.php.

[cutty]

time_zone="-"

display issue.


time_zone="+"  appears ok

[gerbold]

Hello Egon,

I've some doubts regarding the statistics_1 diagram.
   

1) Does the description of the line colors match the values used for the calculation?

There are four lines in the diagram:
- signals (max = n)
- strikes/signals
- involved/strikes
- used/involved

There are two scales at the diagram:
- left hand percentage form 0% to 100%
- right hand signal count from 0 to n

Thus I expect the right scale to be valid for the signals graph amd the left scale with the percentage to be valid for the three other graphs.
At UTC 2020-05-10 18_29 I took a snapshot of the statistics_1 diagram and of the satitons list figures.
Based on the pixels I calculated the percentages in the graph and based on the figures I calculated the expected percentages. They differ greatly.
The percentages sudenly start to match, when I chose different calculations.
valid/total = 11986/20887 = 57% = yellow = 69px/121px = 57%
involved/total = 5440/20887 = 26% = green = 32px/121px = 26%
used/total = 2208/20887 = 11% ~ blue = 14px/121px = 12%

Is this just coincidence or is the legend partially wrong?
yellow is documented as strikes/signals, but it seems to be valid/total
green is documented as invlolved/strikes, but it seems to be invlolved/total
blue is documented as used/involved, but it seems to be used/total

2) There are sudden drops ot the green line to zero. See May 8th 9am, May 9th 1pm and 6pm, May 10th 8pm.
I'd possibly understand this if there would be interference and the station would suddenly send only unusable signals, but there was no sign of a malfunction. 
The most extreme example is at May 8th 9am.
For more than two hours all valid signals have been also regarded as involved signals.
Then in the next moment zero of the valid signals have been involved.
How can this behaviour be explained?

Klaus

[Egon]

Hi all,

I had little time to take care of it, I will look at the problems and check everything.

Regards.
Egpn

[gerbold]

I have another example with the same results at UTC 2020-05-10 07:04
valid/total = 357/521 = 69% = yellow = 83px/121px = 69%
involved/total = 357/521 = 69% = green = 83px/121px = 69%
used/total = 118/521 = 23% = blue = 28px/121px = 23%
   

[Egon]

You are right, the calculations of the curves were not correct. I corrected it, please check.

I'm still working on the time zone problem.

/Egon

[gerbold]

UTC 2020-05-15 05:45
The persentages seem to be ok:
valid/total = 885/1278 = 69% = yellow = 83px/120px = 69%; The diagram uses the deviating name strikes/signals.
involved/valid = 672/885 = 76% = green = 91px/120px = 76%; The diagram uses the deviating name involved/strikes.
used/involved = 178/672 = 26% = blue = 31px/120px = 26%; The diagram uses the identical name used/involved.

The total signal count appears to be too high. This is valid for all samples I took.
max signals = 31618
total = 1278, white = 7px/120px*31618 = 1844

   

The terminology is slightly different between the https://map.blitzortung.org/statistic_1.php diagram, http://de.blitzortung.org/station_list.php and the table view on LMO https://www.lightningmaps.org/blitzortun...n_id=14972:

- "signals" in the diagram corresponds to "total" in the station list table, which corresponds roughly to "Signlas/h" on LMO (a factor of 2 has to be regarded).
Thus "signals" and "total" can be regarded as synonymous.

- "strikes" in the diagram corresponds to "valid" in the station list table. But "Strokes/h" on LMO corresponds to "involved" in the station list table.
The difference between "valid" and "involved" needs clarification. In different context both are named "strike/stroke". At the moment I assume LMO is right.

Here: https://forum.blitzortung.org/showthread...e#pid21010 Cutty explained the sieve which is applied to filter the valid from the total signals by means of some quality criteria.

The server computes a 'noise level' for a specific channel's signal, on the first 25% of the 'impulse group" analysis time window (1000ms)  (beginning 500ms before 'trigger') 
and will REJECT, that is "Not Consider" a signal whose Trigger Voltage is NOT 3 times that computed noise level.  (that channel selected will be coded "valid" on the BLLITZORTUNG.org  Signal images (or at Sferics US   for "Americas Rergions" stations only LIVE SIGNALS ... "A "valid"channel will move into consideration for "locator"

A signal which meets the above defined quality criteria doesn't automatically qualify to be a stroke. The "locator" has to decide whether the signal actually belongs to a stroke. I assume the "involved" signals did pass the "locator" sucessfully - and thus they qualified as strokes.

[Egon]

Hi,

here is a brief explanation of the 4 values on BO that I introduced a few years ago to replace the efficiency index at BO, which was not useful to adjust a station. (my opinion)

(white) Total strikes is the number of signals received on the compting server. This does not have to be the number of signals that the station has sent, since the signals are sent to the computinging servers via a relay server. Here some signals are lost.

(yellow) Valid signals are signals that look like impulses and have a certain noise / signal ratio. A characteristic signal is calculated for each station. If the received signal differs significantly from it, it is not valid. The graphic shows an example. In addition, the time stamp is also adjusted, as the trigger location depends on the signal strength. In the picture you can see three signals, only channel 3 delivers a valid signal. Below are 3 reference signals from the 6 channels of the station. You can check the valid status also with the grafic  signal.php.

(green) Signals with a similar time stamp are grouped together. If there are enough signals in a group, a calculation is initiated with the amount of signals. These signals are the signals involved. A station far away from all other stations will probably never be involved in a calculation because not enough other stations received the same lightning discharges.

(blue) A position calculation is started for each set of valid signals. Not all signals of the set are 100% suitable. Often only the 12 best matched signals are selected from the set. These are the used signals.

All signals that were involved are listed in the signal files. The signals that were used are marked as used signals. So everyone can process and visualize involved / used signals for themselves.

The values in the statistic_1.php graphic do not have to match the values in the station list on BO exact. The station list values are taken from the database, which is updated every minute. The values for the graphic statistic_1.php are taken from a file to relieve the database. However, this file is updated at different times than the database.

/ Egon

   

[gerbold]

Thanks Egon for the explanation. The content of this thread has been condensed in https://docs.lightningmaps.org/doku.php?...statistics.

[Egon]

Hi,

would remove the parameter &background_transparent=1 in the links

"Statistics 1 of station 2656" and "Statistics 2 of station 2656"

It only makes sense if the background is not white.



The signal graphic also contains the valid information next to the channel number.

https://map.blitzortung.org/signal.php?s...ode=signal

   


/Egon

[gerbold]

Thanks Egon, all three changes are done.