JesseWeather with Notifications is LIVE

tonycondon said:
nevermind. i had poorly set up my notifications so that when the ceiling improved for a short while the IFR note was no longer active. but i had set up MVFR and VFR notifications with visibility requirements. Ceiling was Marginal but Vis was good so no notification was sent. then it went back down to IFR so i got a new notification. I had not set it up to send a notification when it was no longer active.
Click to expand...

CJones said:
Ahhh.. The infamous loose nut behind the keyboard problem. Happens all the time...
Click to expand...

PEBCAK
 
ok. now the METAR at KAAO is reporting 1100 Broken. My MVFR notification is supposed to go off when the ceiling is greater than or equal to 1100 and less than or equal to 3000. i dont think this one is me. why haven't I received a text?
 
and now it is reporting 1300 and i got a text. I'm wondering if greater than or equal to is really just greater than in the code?
 
tonycondon said:
ok. now the METAR at KAAO is reporting 1100 Broken. My MVFR notification is supposed to go off when the ceiling is greater than or equal to 1100 and less than or equal to 3000. i dont think this one is me. why haven't I received a text?
Click to expand...

tonycondon said:
and now it is reporting 1300 and i got a text. I'm wondering if greater than or equal to is really just greater than in the code?
Click to expand...

Funny -- I forgot the notification sender is basically the opposite of what it says. It was originally greater than or equal to except the interface didn't say that. I changed the interface to say that and then changed the backend code to use >= and <=. The problem is the back-end code isn't checking to see if the conditions exist as much as it is checking to see that it does NOT exist. So by setting the does not exist check to use <= it became <.

The above is fixed. It should work right now.
 
great! now if the weather would just cooperate we can test it out again!
 
jesse said:
Funny -- I forgot the notification sender is basically the opposite of what it says. It was originally greater than or equal to except the interface didn't say that. I changed the interface to say that and then changed the backend code to use >= and <=. The problem is the back-end code isn't checking to see if the conditions exist as much as it is checking to see that it does NOT exist. So by setting the does not exist check to use <= it became <.

The above is fixed. It should work right now.
Click to expand...

Hooray for negative logic!
 
It appears that you are still disabling notifications after a change has been made. I thought you went to a 3 minute timer on those?
 
tonycondon said:
It appears that you are still disabling notifications after a change has been made. I thought you went to a 3 minute timer on those?
Click to expand...
I thought I did too. Could you tell me which change causes it to go disabled?
 
Jesse, would it be possible to use wind barbs for the wind direction in the charts? The graph isn't all that intuitive, and kinda annoying when the winds are out of the north. The other option which isn't nearly as good would be to normalize the wind scale on the side, by putting the "average" direction in the middle of the scale.
 
vontresc said:
Jesse, would it be possible to use wind barbs for the wind direction in the charts? The graph isn't all that intuitive, and kinda annoying when the winds are out of the north. The other option which isn't nearly as good would be to normalize the wind scale on the side, by putting the "average" direction in the middle of the scale.
Click to expand...
I've messed with doing all kinds of **** for the wind -- but the graphing library sort of sucks. It tends to put things on top of each-other when I try to add images or text to the data points.

I'm thinking about trying to do some fancy on-hover something with javascript.

I could switch to a more powerful flash graphing library but then it wouldn't work on the iPhone.
 
jesse said:
I've messed with doing all kinds of **** for the wind -- but the graphing library sort of sucks. It tends to put things on top of each-other when I try to add images or text to the data points.

I'm thinking about trying to do some fancy on-hover something with javascript.

I could switch to a more powerful flash graphing library but then it wouldn't work on the iPhone.
Click to expand...

What about something like this, Jesse? Y-scale would be wind velocity, zero in the middle, North is up and South is down. X-scale would be time. I suppose it could get messy with 090 and 270, but it would show trends in both strength and wind direction change nicely.

attachment.php


Got this from a MathWorks page on various methods to graph winds.
 

Attachments

  • ScreenShot040.jpg
    ScreenShot040.jpg
    15.6 KB · Views: 122
TangoWhiskey said:
What about something like this, Jesse? Y-scale would be wind velocity, zero in the middle, North is up and South is down. X-scale would be time. I suppose it could get messy with 090 and 270, but it would show trends in both strength and wind direction change nicely.

attachment.php


Got this from a MathWorks page on various methods to graph winds.
Click to expand...
Well I don't think my graphing library would support that. I'm pretty much limited to that which is possible here:
http://hem.bredband.net/aditus/chunkhtml/index.html

The big problem with something like that is the fact that hte graphs are dynamic. If you were doing a graph for some presentation or something that'd be fine. A graph that gets auto-generated based on completely random conditions and time frames would likely have all of those lines on top of each-other in every which direction.
 
Another idea; use solid colored dots to plot the time; dark green would be "current", you could fade to light green going back in time and use another color (dark yellow going to light yellow) for the TAF future estimates. You wouldn't have as many data points as this one does, so it wouldn't look as "busy":

attachment.php


Source: USGS Wind Rose Data Grapher

Edit: Posted this before I saw your list of "what is possible" with the library you're using.
 

Attachments

  • ScreenShot041.jpg
    ScreenShot041.jpg
    87.5 KB · Views: 124
jesse said:
If you pay for the professional version. I can't see how you could show "time" with something like that either.
Click to expand...

Ah, that "pay" thing again. ;-)

I'd just line up a row or two of those, each one plotting just one wind point in time, each row covering six graphs / six hours. Each one labeled for which time it represents. Couldn't overlay it with any other graphs, of course.
 
Anyone willing to help with the writing of a press release? At 33 members right now. Also -- I'm thinking some pro pilot forums might be a good place to mention it. Does anyone know of an active pro pilot forum that has some sort of advertising / classifieds forum?
 
Last edited:
jesse said:
I thought I did too. Could you tell me which change causes it to go disabled?
Click to expand...

it appears that i deleted ALL of the notification requirements, and then added some more requirements back to it. As soon as all of the requirements were deleted the notification was disabled, and adding new requirements did not enable it.
 
jesse said:
Anyone willing to help with the writing of a press release? At 33 members right now. Also -- I'm thinking some pro pilot forums might be a good place to mention it. Does anyone know of an active pro pilot forum that has some sort of advertising / classifieds forum?
Click to expand...

I'd be happy to give the press release a try...I'll write something up this weekend and let you have a read...
 
i don't know about anyone else, but queries to the site are not working for me today, and notifications are not going out either.
 
Either the National Weather Service blocked me -- or they are having technical problems. I'm looking into it.
 
fwiw this is what i get. you probably already knew this

attachment.php
 

Attachments

  • jesse.bmp
    2.2 MB · Views: 109
Someone made some firewall changes that caused my connection to the NOAA to break. It should be fixed soon.
 
i just got a notification so it must be working!
 
It's fixed. There may be a gap in the data. I'll look into that later...Sorry
 
suggestion:

sometimes the info as graphed can be decieving because of few or scattered layers. this makes it look like the ceiling may be better than it really is. of course i know that few and scattered layers are not technically ceilings but i also know that often they "fill in the gap" between good and bad weather. perhaps you could graph few and scattered layers with a different color?

for example lets say the weather is few 1000 and broken 2000. the graph would be white at the ground going up to 1000 feet, then maybe it would be yellow from 1000 to 2000 then red from 2000 and above. colors can be whatever you choose. that way you can still see from the graph that there is a cloud layer there, and also know that the actual ceiling is higher. could be really handy...
 
That looks like a good idea, show trend from SCT to BKN.
 
I've thought about trying to show it somehow Tony. The color method you describe might work well -- I'll have to play with it. The tricky part is what you do when those conditions suddenly don't exist in the next metar and then the graph looks weird.

It may work though. I'll give it a shot.
 
i just pulled the last 30 observations from ICT and made this nifty graph. i basically used 3 colums: Time, Few/Sct, Bkn/Ovc. If there was no Few/Sct layer then i put the Bkn/Ovc height in both columns. Then i graphed it and put the few/Sct graph above the Bkn/Ovc graph. Green is clear air, Yellow is Few/Sct and Orange (red was too bright) is the ceiling.

attachment.php
 

Attachments

  • wx.bmp
    1.4 MB · Views: 298
some more feedback,

the TAF display doesnt consider the effective time of the forecast, it just plots each "division time" within the forecast. For this TAF:

KICT 131140Z 1312/1412 05013KT 4SM BR OVC008
TEMPO 1312/1315 2SM -DZ BR OVC004
FM131500 08013KT 3SM -DZ BR OVC005
FM131800 08013KT 4SM -RA BR OVC008

Nothing gets plotted between 1800 and 1200 (a long time!). The 800 Overcast forecast should be extended out for that period.
 
tonycondon said:
some more feedback,

the TAF display doesnt consider the effective time of the forecast, it just plots each "division time" within the forecast. For this TAF:

KICT 131140Z 1312/1412 05013KT 4SM BR OVC008
TEMPO 1312/1315 2SM -DZ BR OVC004
FM131500 08013KT 3SM -DZ BR OVC005
FM131800 08013KT 4SM -RA BR OVC008

Nothing gets plotted between 1800 and 1200 (a long time!). The 800 Overcast forecast should be extended out for that period.
Click to expand...


That's good feedback right there.
 
tonycondon said:
some more feedback,

the TAF display doesnt consider the effective time of the forecast, it just plots each "division time" within the forecast. For this TAF:

KICT 131140Z 1312/1412 05013KT 4SM BR OVC008
TEMPO 1312/1315 2SM -DZ BR OVC004
FM131500 08013KT 3SM -DZ BR OVC005
FM131800 08013KT 4SM -RA BR OVC008

Nothing gets plotted between 1800 and 1200 (a long time!). The 800 Overcast forecast should be extended out for that period.
Click to expand...
Yeah. That's pretty retarded. I guess the guy that wrote it always refused to read TAFs stating the format was stupid. That same guy decides to write code to interrpet TAFs and botches it.

The TAF time processing stuff is all pretty screwy..mainly because when I wrote it..I barely knew how to read one. I'll get it fixed.

Good thing that it is Beta :D
 
good thing im the 'typical idiot user' who blunders into all the stuff that could be better...
 
different kind of notification today
KE38 281505Z AUTO 22029G44KT 10SM CLR 21/06 A2963
 
tonycondon said:
i just pulled the last 30 observations from ICT and made this nifty graph. i basically used 3 colums: Time, Few/Sct, Bkn/Ovc. If there was no Few/Sct layer then i put the Bkn/Ovc height in both columns. Then i graphed it and put the few/Sct graph above the Bkn/Ovc graph. Green is clear air, Yellow is Few/Sct and Orange (red was too bright) is the ceiling.

attachment.php
Click to expand...

That looks more like a messed up candy corn.:D
 
Hey neat, I just got my first real notification. Its snowing and low ceilings in Albuquerque. Great job, Jesse!
 
You must log in or register to reply here.
Back
Top