Pressure with trend
Re: Pressure with trend
Hi Holli!
Why do you want a small pressure change to be 30 degrees and a big to be 45 degrees!
I still want it to be proportional and also with a limit that is reached now and then. Then
it's easy to see when the weather is changing.
In other topic I have suggested to introduce "YoWindow Weather Alerts" and this could
be used when the limit is reached.
Why do you want a small pressure change to be 30 degrees and a big to be 45 degrees!
I still want it to be proportional and also with a limit that is reached now and then. Then
it's easy to see when the weather is changing.
In other topic I have suggested to introduce "YoWindow Weather Alerts" and this could
be used when the limit is reached.
Re: Pressure with trend
Due to its effects. Even a small change indicates a high or a low, but a great change is not proportional to the upcoming winds or precipation.daddylito wrote:Hi Holli!
Why do you want a small pressure change to be 30 degrees and a big to be 45 degrees!
I don't think much of alerts depending on the values of a single location. If there is a chance to take weather alerts from an official weather service, they should be taken, but not the values of a local station. Alerts based on such simple parameters are in more than the half of all cases wrong.In other topic I have suggested to introduce "YoWindow Weather Alerts" and this could
be used when the limit is reached.
Re: Pressure with trend
Hi again!
You misunderstand what I mean with an alert.
I only mean a simple message on the screen
e.g. "Pressure has dropped 7 hPa the last 3 hours",
and no weather data evaluation of the risk of a coming storm.
You misunderstand what I mean with an alert.
I only mean a simple message on the screen
e.g. "Pressure has dropped 7 hPa the last 3 hours",
and no weather data evaluation of the risk of a coming storm.
Re: Pressure with trend
This could be signalized by coloring the arrow. A nice feature would be a hover information in a bubble while the mouse is hovering on the pressure.
- par
- Posts: 8724
- Joined: Mon Sep 21, 2009 11:56 am
- Name: Pasha
- Location: Saint-Petersburg, Russia
- Contact:
Re: Pressure with trend
You may test the pressure-trend feature in the latest release-candidate.
Now there is an arrow in the Inspector beside the current pressure value.
The arrow indicates the pressure trend.
I liked Holli explanation of the arctangent approach, I have used it.
There is no limit on trend value.
1hPa/3h gives 15 deegrees.
3hPa/3h gives 45 degrees.
9hPa/3h gives 65 degrees.
When the trend is greater than 4.5hPa/3h, the arrow gets colored in red indicating rapid pressure change.
When you move the mouse over the trend arrow the trend value displayed with a verbal comment "raising, falling or steady".
Just like Holli have proposed.
There is no support for PWS stations yet, please wait a bit.
Despite everything seems to work there are several questions I need to clear up with you to finish the pressure-trend feature.
For the start, let me share with you the internals of pressure trend picking algorithm.
METAR stations provide the weather in a certain time intervals.
Depending on the station the interval varies between 30 minutes to 2 hours.
YoWindow keeps these weather measurements in history for 6 hours.
I had to choose the time interval to compute the trend for.
Now the interval is 3 hours.
The inteval is the time difference between the last measurement and the time 3 hours before it.
I call this time - the perfect-time.
perfect-time = last-observation-time - 3 hours
In 99% cases there are no data on pressure at the perfect-time.
I'm guessing pressure at the perfect-time by interpolating between previous and next weather measurements from history.
Then the trend (hPa/hour) is computed as follows
trend = (observed-pressure - perfect-time-pressure) / 3
That's the algorithm.
Question 1.
If I have weather history for the last hour but do not have older weather, would it be correct to provide the trend
for 1 hour interval?
Or it would be better to NOT provide the trend until the required weather history is accumulated.
What do you think is the minimal pressure trend interval I can count as a valid one?
30 minutes?
1 hour?
2 hours?
3 hours?
Question 2.
Does it make sense to display pressure trend for the weather forecast?
The weather is provided in 6 hour intervals.
Some intervals are 12 hour long.
Question 3.
Many weather services and programs indicate the trend by one of three states: raising, falling, steady.
I want YoWindow to indicate pressure this way in the tooltip (in addition to the arrow).
And I want YoWindow to use the common rules for this verbal trend indication.
Please, help me to find the value when steady trend turnes into raising one.
Now there is an arrow in the Inspector beside the current pressure value.
The arrow indicates the pressure trend.
I liked Holli explanation of the arctangent approach, I have used it.
There is no limit on trend value.
1hPa/3h gives 15 deegrees.
3hPa/3h gives 45 degrees.
9hPa/3h gives 65 degrees.
When the trend is greater than 4.5hPa/3h, the arrow gets colored in red indicating rapid pressure change.
When you move the mouse over the trend arrow the trend value displayed with a verbal comment "raising, falling or steady".
Just like Holli have proposed.
There is no support for PWS stations yet, please wait a bit.
Despite everything seems to work there are several questions I need to clear up with you to finish the pressure-trend feature.
For the start, let me share with you the internals of pressure trend picking algorithm.
METAR stations provide the weather in a certain time intervals.
Depending on the station the interval varies between 30 minutes to 2 hours.
YoWindow keeps these weather measurements in history for 6 hours.
I had to choose the time interval to compute the trend for.
Now the interval is 3 hours.
The inteval is the time difference between the last measurement and the time 3 hours before it.
I call this time - the perfect-time.
perfect-time = last-observation-time - 3 hours
In 99% cases there are no data on pressure at the perfect-time.
I'm guessing pressure at the perfect-time by interpolating between previous and next weather measurements from history.
Then the trend (hPa/hour) is computed as follows
trend = (observed-pressure - perfect-time-pressure) / 3
That's the algorithm.
Question 1.
If I have weather history for the last hour but do not have older weather, would it be correct to provide the trend
for 1 hour interval?
Or it would be better to NOT provide the trend until the required weather history is accumulated.
What do you think is the minimal pressure trend interval I can count as a valid one?
30 minutes?
1 hour?
2 hours?
3 hours?
Question 2.
Does it make sense to display pressure trend for the weather forecast?
The weather is provided in 6 hour intervals.
Some intervals are 12 hour long.
Question 3.
Many weather services and programs indicate the trend by one of three states: raising, falling, steady.
I want YoWindow to indicate pressure this way in the tooltip (in addition to the arrow).
And I want YoWindow to use the common rules for this verbal trend indication.
Please, help me to find the value when steady trend turnes into raising one.
Re: Pressure with trend
Hi Pasha!
Thanks for your great and fast work concerning this feature!
I have had a busy day,
but since I by mail promised you an answer today,
I send you these comments!
Wish you a pleasant weekend
/Daddy
My comments:
You have made 3 questions, but my first comment concerns
the display of the value of pressure change.
- I like it to be an arrow
- I like the arrow to be red when a limit is reached
- I like the hoover info, i.e. when pointing your mouse you get the actual value in digits
- The hoover info should have a decimal digit in order to see the difference of 0.6 and 1.4
- I still like a proportional solution. I like to see a small change while the pressure changes with a medium rate,
and not only a small change when it´s near to stable
- And there is no problem for me that the trend value arrow could have a limit,
as long as hoovering info can give the actual value. The angle range of the arrow is better used!
Q1) What is the minimal pressure trend interval I can count as a valid one?
- Sorry, I can't give you any values! Someone else, please help!
Q2) Does it make sense to display pressure trend for the weather forecast?
- It is a prediction of whats going to happen, so I say yes.
Q3) The value when steady trend turns into raising or falling.
- Sorry, I can't give you any values! Someone else, please help!
Thanks for your great and fast work concerning this feature!
I have had a busy day,
but since I by mail promised you an answer today,
I send you these comments!
Wish you a pleasant weekend
/Daddy
My comments:
You have made 3 questions, but my first comment concerns
the display of the value of pressure change.
- I like it to be an arrow
- I like the arrow to be red when a limit is reached
- I like the hoover info, i.e. when pointing your mouse you get the actual value in digits
- The hoover info should have a decimal digit in order to see the difference of 0.6 and 1.4
- I still like a proportional solution. I like to see a small change while the pressure changes with a medium rate,
and not only a small change when it´s near to stable
- And there is no problem for me that the trend value arrow could have a limit,
as long as hoovering info can give the actual value. The angle range of the arrow is better used!
Q1) What is the minimal pressure trend interval I can count as a valid one?
- Sorry, I can't give you any values! Someone else, please help!
Q2) Does it make sense to display pressure trend for the weather forecast?
- It is a prediction of whats going to happen, so I say yes.
Q3) The value when steady trend turns into raising or falling.
- Sorry, I can't give you any values! Someone else, please help!
- par
- Posts: 8724
- Joined: Mon Sep 21, 2009 11:56 am
- Name: Pasha
- Location: Saint-Petersburg, Russia
- Contact:
Re: Pressure with trend
Thanks, Daddy!
Holli where are you?
Holli where are you?
Re: Pressure with trend
Pasha, here are the information about the pressure trend as used in one of the JavaScript (for the Weather Display software):
- the original (Ken True) trend words: Steady, Rising Slowly, Rising Rapidly, Falling Slowly, Falling Rapidly
- the original (Ken True) JavaScript code associated with those words:
The above code was designed using the following rate of change information:
But I was not satisfied with the trend results so I made some changes to the pressure trend (added 2 trend levels):
- Ray's new trend words: Steady, Gently Rising, Rising Slowly, Rising Rapidly, Gently Falling, Falling Slowly, Falling Rapidly
- Ray's JavaScript code associated with those words:
My code was redesigned using the following rate of change information:
The pressure trend numeric value, which can be positive, zero, or negative is produced by Weather Display and is based on a 3-hour pressure change - the trend value is in hPa or millibars.
The JavaScript also produces 2 icons:
- a green up arrow
- a red down arrow
Obviously, I had to modify that limited display, so I added an additional icon:
- a green up arrow - pressure going up
- a red and green horizontal bar - pressure steady (not changing)
- a red down arrow - pressure going down
Now to answer your questions...
Question 1: If you have only the pressure trend for the past hour, you should include that value until it can be replaced with the pressure trend for the past 3 hours.
Question 2: I am not too sure I understand what you meant when you wrote "Does it make sense to display pressure trend for the weather forecast?". The pressure trend is a very big part of the immediate forecast and can tell what is happening even if a new and complete forecast from some other source is only provided every 6 to 12 hours. For example, if there is a big drop of pressure, then I can expect some really bad weather and I do not have to be told every 6 to 12 hours that bad weather is coming, I can see it by just looking at the pressure change.
Question 3: Pasha, remember the expression "KISS" (almost meaning keep it simply simple ...
). Only 3 icons should be used for the 3 states: raising, falling, and steady. I think and feel that the arctangent approach might and will make thing too complicated for the average user. Again, "KISS"...
- the original (Ken True) trend words: Steady, Rising Slowly, Rising Rapidly, Falling Slowly, Falling Rapidly
- the original (Ken True) JavaScript code associated with those words:
Code: Select all
if ((btrnd >= -0.7) && (btrnd <= 0.7)) { return(langBaroTrend[0]); } // Steady
if ((btrnd > 0.7) && (btrnd < 2.0)) { return(langBaroTrend[1]); } // Rising Slowly
if (btrnd >= 2.0) { return(langBaroTrend[2]); } // Rising Rapidly
if ((btrnd < -0.7) && (btrnd > -2.0)) { return(langBaroTrend[3]); } // Falling Slowly
if (btrnd <= -2.0) { return(langBaroTrend[4]); } // Falling Rapidly
return(btrnd);
Code: Select all
Change Rates
Rapidly =.06 inHg; 1.5 mm Hg; 2 hPa; 2 mb
Slowly =.02 inHg; 0.5 mm Hg; 0.7 hPa; 0.7 mb
- Ray's new trend words: Steady, Gently Rising, Rising Slowly, Rising Rapidly, Gently Falling, Falling Slowly, Falling Rapidly
- Ray's JavaScript code associated with those words:
Code: Select all
switch (true) {
case ((btrnd == 0.0) || (btrnd == -0.0)) : barotrend = langBaroTrend[0]; break; // Steady
case ((btrnd > 0.0) && (btrnd < 0.7)) : barotrend = langBaroTrend[1]; break; // Gently Rising
case ((btrnd >= 0.7) && (btrnd < 2.0)) : barotrend = langBaroTrend[2]; break; // Rising Slowly
case ((btrnd >= 2.0)) : barotrend = langBaroTrend[3]; break; // Rising Rapidly
case ((btrnd < -0.0) && (btrnd > -0.7)) : barotrend = langBaroTrend[4]; break; // Gently Falling
case ((btrnd <= -0.7) && (btrnd > -2.0)) : barotrend = langBaroTrend[5]; break; // Falling Slowly
case ((btrnd <= -2.0)) : barotrend = langBaroTrend[6]; break; // Falling Rapidly
default : barotrend = langBaroTrend[0]; break; // Steady
}
return(barotrend);
Code: Select all
Change Rates
Rapidly = 2 hPa; 2 mb; 0.06 inHg; 1.5 mm Hg;
Slowly = 0.7 hPa; 0.7 mb; 0.02 inHg; 0.5 mm Hg
Gently = < 0.7 hPa; < 0.7 mb; < 0.02 inHg; < 0.5 mm Hg
The JavaScript also produces 2 icons:
- a green up arrow
- a red down arrow
Obviously, I had to modify that limited display, so I added an additional icon:
- a green up arrow - pressure going up
- a red and green horizontal bar - pressure steady (not changing)
- a red down arrow - pressure going down
Now to answer your questions...
Question 1: If you have only the pressure trend for the past hour, you should include that value until it can be replaced with the pressure trend for the past 3 hours.
Question 2: I am not too sure I understand what you meant when you wrote "Does it make sense to display pressure trend for the weather forecast?". The pressure trend is a very big part of the immediate forecast and can tell what is happening even if a new and complete forecast from some other source is only provided every 6 to 12 hours. For example, if there is a big drop of pressure, then I can expect some really bad weather and I do not have to be told every 6 to 12 hours that bad weather is coming, I can see it by just looking at the pressure change.
Question 3: Pasha, remember the expression "KISS" (almost meaning keep it simply simple ...

Do you think any of the information provided above could be useful?Please, help me to find the value when steady trend turnes into raising one.
- par
- Posts: 8724
- Joined: Mon Sep 21, 2009 11:56 am
- Name: Pasha
- Location: Saint-Petersburg, Russia
- Contact:
Re: Pressure with trend
That is the information I was looking for,
Ray, thank you much!
These are the changes to the pressure trend implementation.
1. The arrow angle reflects the trend in 5 fixed positions.
0° - steady
25° - raising, trend > 0.7 hPa / 3 hour
70° - rapidly raising, trend > 2 hPa / 3 hour, the arrow is colored in red
-25° - falling
-70° - rapidly falling
I have abandoned the idea of smooth arrow rotation since it may confuse the user.
It seems to me, that reliable indication of pressure behavior (steady/changing/rapidly changing) is more important.
2. In case there is no recorded pressure history older than 1 hour, the pressure trend is not provided.
Ray, thanks again!
3. The trend value is now displayed in the tooltip with a 1 decimal digit precision.
Daddy, thanks for pointing at this error.
4. PWS supprot is coming soon.
It seems I have finished the feature now.
Thanks everyone!
You may try it out.
YoWindow is a full-featured barometer now
http://repkasoft.com/rc.html
Ray, thank you much!
These are the changes to the pressure trend implementation.
1. The arrow angle reflects the trend in 5 fixed positions.
0° - steady
25° - raising, trend > 0.7 hPa / 3 hour
70° - rapidly raising, trend > 2 hPa / 3 hour, the arrow is colored in red
-25° - falling
-70° - rapidly falling
I have abandoned the idea of smooth arrow rotation since it may confuse the user.
It seems to me, that reliable indication of pressure behavior (steady/changing/rapidly changing) is more important.
2. In case there is no recorded pressure history older than 1 hour, the pressure trend is not provided.
Ray, thanks again!
3. The trend value is now displayed in the tooltip with a 1 decimal digit precision.
Daddy, thanks for pointing at this error.
4. PWS supprot is coming soon.
It seems I have finished the feature now.
Thanks everyone!
You may try it out.
YoWindow is a full-featured barometer now

http://repkasoft.com/rc.html
-
- Posts: 1645
- Joined: Wed Nov 18, 2009 9:58 am
- Location: Vienna, Austria
Re: Pressure with trend
Ähemm - i don't know if you intended to display the arrow in the system-tray-window too.par wrote:4. PWS supprot is coming soon.
If so, it's missing

If not: perfect!
»Man is impressive. In a universe full of untold secrets he has invented boredom.«