[Met_help] [rt.rap.ucar.edu #93208] History for METViewer v2.11 question
John Halley Gotway via RT
met_help at ucar.edu
Thu Nov 21 15:12:32 MST 2019
----------------------------------------------------------------
Initial Request
----------------------------------------------------------------
When plotting continuous statistics ME and RMSE using MET input .stat files from Point-Stat, which settings in the "Statistics" control panel of the web GUI would generate statistics which match those produced in the MET text output file CNT.txt?
Thanks.
R/
John
Mr John W. Raby, Meteorologist
U.S. Army Research Laboratory
White Sands Missile Range, NM 88002
(575) 678-2004 DSN 258-2004
FAX (575) 678-1230 DSN 258-1230
Email: john.w.raby2.civ at mail.mil<mailto:john.raby at us.army.mil>
----------------------------------------------------------------
Complete Ticket History
----------------------------------------------------------------
Subject: METViewer v2.11 question
From: John Halley Gotway
Time: Thu Nov 21 09:38:36 2019
John,
I see that you have a question about how the stats computed by
METviewer
compare to those computed by MET.
Let's assume that when you ran Point-Stat, you requested the output
line
types SL1L2 and CNT. When you load those .stat files into METviewer,
you
can select specific line types to load, or you can load all of them.
I'll
assume that you loaded all of them into METviewer.
METviewer supports 3 options in the "Statistics" panel. Below is a
description of how the MET output is used for each of these options:
(1) Summary: Pull the statistics (e.g. ME or RMSE) from the MET CNT
line
type directly from the database. For each independent variable value
(i.e.
value on the x-axis), plot either the "Median", "Mean", or "Sum" of
those
statistics based on how the "Plot statistic" dropdown option is set.
(2) Aggregation Statistics: When processing continuous statistics, the
dropdown list below should be set to "Scalar partial sums (SL1L2)" (or
SAL1L2 for anomaly correlation). Rather than selecting the CNT
statistics
pre-computed by MET, select the partial sums from the SL1L2 line type.
For
each independent variable value (i.e. value on the x-axis), aggregate
the
corresponding SL1L2 partial sums together by computing a weighted
average
where the weight is the number of matched pairs (i.e. TOTAL column)
for
each line. From those aggregate partial sums, compute and plot a
single
statistic (e.g. ME or RMSE).
(3) Revision Statistics: This is relatively new. I'm not sure whether
we
get the statistics using method (1) or (2) above. But once we have
them,
we sort them in time and compute "revisions", meaning the change in
statistic value from one time to the next. And it's those revisions
which
are plotted.
To put is simply, method (1) plots the mean of daily statistics while
method (2) plots a single aggregated statistic. Method (1) is
plotting the
statistics reported directly by Point-Stat.
Thanks,
John
On Thu, Nov 21, 2019 at 8:10 AM Raby, John W USA CIV via RT <
met_help at ucar.edu> wrote:
>
> Thu Nov 21 08:09:37 2019: Request 93208 was acted upon.
> Transaction: Ticket created by john.w.raby2.civ at mail.mil
> Queue: met_help
> Subject: METViewer v2.11 question
> Owner: Nobody
> Requestors: john.w.raby2.civ at mail.mil
> Status: new
> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=93208 >
>
>
> When plotting continuous statistics ME and RMSE using MET input
.stat
> files from Point-Stat, which settings in the "Statistics" control
panel of
> the web GUI would generate statistics which match those produced in
the MET
> text output file CNT.txt?
>
> Thanks.
>
> R/
> John
>
> Mr John W. Raby, Meteorologist
>
> U.S. Army Research Laboratory
>
> White Sands Missile Range, NM 88002
>
> (575) 678-2004 DSN 258-2004
>
> FAX (575) 678-1230 DSN 258-1230
>
> Email: john.w.raby2.civ at mail.mil<mailto:john.raby at us.army.mil>
>
>
------------------------------------------------
Subject: METViewer v2.11 question
From: Raby, John W USA CIV
Time: Thu Nov 21 10:28:25 2019
John -
Thanks for this very helpful explanation. In my case, You are correct
in your assumption, I ran Point-Stat specifying output line types
SL1L2 and CNT plus some other line types as well as the .stat output.
When I loaded the database, I specified a folder containing the
MPR.txt files and the .stat files. See attached load_spec_file. When
visualizing in METViewer, I specify STAT for the plot data. I used
option (1) for my plotting which generated plots of ME and RMSE as a
function of lead time. Based on your info below, the series of ME and
RMSE values plotted will come from the CNT line type data in the .stat
files. I noticed that when I plot using the "Mean" setting the plot
looks different than when using the "Median" setting which is
expected. Which of these two settings will produce the ME and RMSE
values which appear in the CNT.txt file?
By specifying the <load_mpr>true</load_mpr> in the load_spec_file as
"true", does that load the MPR line type data from the .stat file or
the MPR.txt files? I guess my question is what does this line in the
load_spec_file do? As you will see in the attached load_spec_file, I
also specify <load_stat>true</load_stat>, which I assume loads the
database with the data from the .stat file. I don't understand why
there would be the option of specifying the loading of the MPR files
separately when the MPR line type data is present in the .stat file.
R/
John
________________________________________
From: John Halley Gotway via RT [met_help at ucar.edu]
Sent: Thursday, November 21, 2019 9:38 AM
To: Raby, John W CIV USARMY CCDC ARL (USA)
Subject: [Non-DoD Source] Re: [rt.rap.ucar.edu #93208] METViewer v2.11
question
All active links contained in this email were disabled. Please verify
the identity of the sender, and confirm the authenticity of all links
contained within the message prior to copying and pasting the address
to a Web browser.
----
John,
I see that you have a question about how the stats computed by
METviewer
compare to those computed by MET.
Let's assume that when you ran Point-Stat, you requested the output
line
types SL1L2 and CNT. When you load those .stat files into METviewer,
you
can select specific line types to load, or you can load all of them.
I'll
assume that you loaded all of them into METviewer.
METviewer supports 3 options in the "Statistics" panel. Below is a
description of how the MET output is used for each of these options:
(1) Summary: Pull the statistics (e.g. ME or RMSE) from the MET CNT
line
type directly from the database. For each independent variable value
(i.e.
value on the x-axis), plot either the "Median", "Mean", or "Sum" of
those
statistics based on how the "Plot statistic" dropdown option is set.
(2) Aggregation Statistics: When processing continuous statistics, the
dropdown list below should be set to "Scalar partial sums (SL1L2)" (or
SAL1L2 for anomaly correlation). Rather than selecting the CNT
statistics
pre-computed by MET, select the partial sums from the SL1L2 line type.
For
each independent variable value (i.e. value on the x-axis), aggregate
the
corresponding SL1L2 partial sums together by computing a weighted
average
where the weight is the number of matched pairs (i.e. TOTAL column)
for
each line. From those aggregate partial sums, compute and plot a
single
statistic (e.g. ME or RMSE).
(3) Revision Statistics: This is relatively new. I'm not sure whether
we
get the statistics using method (1) or (2) above. But once we have
them,
we sort them in time and compute "revisions", meaning the change in
statistic value from one time to the next. And it's those revisions
which
are plotted.
To put is simply, method (1) plots the mean of daily statistics while
method (2) plots a single aggregated statistic. Method (1) is
plotting the
statistics reported directly by Point-Stat.
Thanks,
John
On Thu, Nov 21, 2019 at 8:10 AM Raby, John W USA CIV via RT <
met_help at ucar.edu> wrote:
>
> Thu Nov 21 08:09:37 2019: Request 93208 was acted upon.
> Transaction: Ticket created by john.w.raby2.civ at mail.mil
> Queue: met_help
> Subject: METViewer v2.11 question
> Owner: Nobody
> Requestors: john.w.raby2.civ at mail.mil
> Status: new
> Ticket <Caution-url: Caution-
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=93208 >
>
>
> When plotting continuous statistics ME and RMSE using MET input
.stat
> files from Point-Stat, which settings in the "Statistics" control
panel of
> the web GUI would generate statistics which match those produced in
the MET
> text output file CNT.txt?
>
> Thanks.
>
> R/
> John
>
> Mr John W. Raby, Meteorologist
>
> U.S. Army Research Laboratory
>
> White Sands Missile Range, NM 88002
>
> (575) 678-2004 DSN 258-2004
>
> FAX (575) 678-1230 DSN 258-1230
>
> Email: john.w.raby2.civ at mail.mil<Caution-
mailto:john.raby at us.army.mil>
>
>
------------------------------------------------
Subject: METViewer v2.11 question
From: John Halley Gotway
Time: Thu Nov 21 11:00:27 2019
John,
To help clarify what's going on, please go to the "Common" tab and
click
the checkbox for "Display Number of Stats". Then regenerate your plot
and
look at the top axis. For each independent variable value, this tells
you
the number of stats represented by the dot that is plotted for that
lead
time.
When plotting the mean or median of those stats, it makes sense that
the
summary value changes. But neither of those are reported by MET
directly... since Point-Stat is run once for each vx time. If you'd
like
to see the stats computed directly by Point-Stat, I'd recommend
switching
from the "Series" plot type to the "Box" plot type (for boxplots).
John
On Thu, Nov 21, 2019 at 10:28 AM Raby, John W USA CIV via RT <
met_help at ucar.edu> wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=93208 >
>
> John -
>
> Thanks for this very helpful explanation. In my case, You are
correct in
> your assumption, I ran Point-Stat specifying output line types SL1L2
and
> CNT plus some other line types as well as the .stat output. When I
loaded
> the database, I specified a folder containing the MPR.txt files and
the
> .stat files. See attached load_spec_file. When visualizing in
METViewer, I
> specify STAT for the plot data. I used option (1) for my plotting
which
> generated plots of ME and RMSE as a function of lead time. Based on
your
> info below, the series of ME and RMSE values plotted will come from
the CNT
> line type data in the .stat files. I noticed that when I plot using
the
> "Mean" setting the plot looks different than when using the "Median"
> setting which is expected. Which of these two settings will produce
the ME
> and RMSE values which appear in the CNT.txt file?
>
> By specifying the <load_mpr>true</load_mpr> in the load_spec_file as
> "true", does that load the MPR line type data from the .stat file or
the
> MPR.txt files? I guess my question is what does this line in the
> load_spec_file do? As you will see in the attached load_spec_file, I
also
> specify <load_stat>true</load_stat>, which I assume loads the
database with
> the data from the .stat file. I don't understand why there would be
the
> option of specifying the loading of the MPR files separately when
the MPR
> line type data is present in the .stat file.
>
> R/
> John
>
>
> ________________________________________
> From: John Halley Gotway via RT [met_help at ucar.edu]
> Sent: Thursday, November 21, 2019 9:38 AM
> To: Raby, John W CIV USARMY CCDC ARL (USA)
> Subject: [Non-DoD Source] Re: [rt.rap.ucar.edu #93208] METViewer
v2.11
> question
>
> All active links contained in this email were disabled. Please
verify the
> identity of the sender, and confirm the authenticity of all links
contained
> within the message prior to copying and pasting the address to a Web
> browser.
>
>
>
>
> ----
>
> John,
>
> I see that you have a question about how the stats computed by
METviewer
> compare to those computed by MET.
>
> Let's assume that when you ran Point-Stat, you requested the output
line
> types SL1L2 and CNT. When you load those .stat files into
METviewer, you
> can select specific line types to load, or you can load all of them.
I'll
> assume that you loaded all of them into METviewer.
>
> METviewer supports 3 options in the "Statistics" panel. Below is a
> description of how the MET output is used for each of these options:
>
> (1) Summary: Pull the statistics (e.g. ME or RMSE) from the MET CNT
line
> type directly from the database. For each independent variable
value (i.e.
> value on the x-axis), plot either the "Median", "Mean", or "Sum" of
those
> statistics based on how the "Plot statistic" dropdown option is set.
>
> (2) Aggregation Statistics: When processing continuous statistics,
the
> dropdown list below should be set to "Scalar partial sums (SL1L2)"
(or
> SAL1L2 for anomaly correlation). Rather than selecting the CNT
statistics
> pre-computed by MET, select the partial sums from the SL1L2 line
type. For
> each independent variable value (i.e. value on the x-axis),
aggregate the
> corresponding SL1L2 partial sums together by computing a weighted
average
> where the weight is the number of matched pairs (i.e. TOTAL column)
for
> each line. From those aggregate partial sums, compute and plot a
single
> statistic (e.g. ME or RMSE).
>
> (3) Revision Statistics: This is relatively new. I'm not sure
whether we
> get the statistics using method (1) or (2) above. But once we have
them,
> we sort them in time and compute "revisions", meaning the change in
> statistic value from one time to the next. And it's those revisions
which
> are plotted.
>
> To put is simply, method (1) plots the mean of daily statistics
while
> method (2) plots a single aggregated statistic. Method (1) is
plotting the
> statistics reported directly by Point-Stat.
>
> Thanks,
> John
>
> On Thu, Nov 21, 2019 at 8:10 AM Raby, John W USA CIV via RT <
> met_help at ucar.edu> wrote:
>
> >
> > Thu Nov 21 08:09:37 2019: Request 93208 was acted upon.
> > Transaction: Ticket created by john.w.raby2.civ at mail.mil
> > Queue: met_help
> > Subject: METViewer v2.11 question
> > Owner: Nobody
> > Requestors: john.w.raby2.civ at mail.mil
> > Status: new
> > Ticket <Caution-url: Caution-
> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=93208 >
> >
> >
> > When plotting continuous statistics ME and RMSE using MET input
.stat
> > files from Point-Stat, which settings in the "Statistics" control
panel
> of
> > the web GUI would generate statistics which match those produced
in the
> MET
> > text output file CNT.txt?
> >
> > Thanks.
> >
> > R/
> > John
> >
> > Mr John W. Raby, Meteorologist
> >
> > U.S. Army Research Laboratory
> >
> > White Sands Missile Range, NM 88002
> >
> > (575) 678-2004 DSN 258-2004
> >
> > FAX (575) 678-1230 DSN 258-1230
> >
> > Email: john.w.raby2.civ at mail.mil<Caution-
mailto:john.raby at us.army.mil>
> >
> >
>
>
>
------------------------------------------------
Subject: RE: [Non-DoD Source] Re: [rt.rap.ucar.edu #93208] METViewer v2.11 question
From: Raby, John W USA CIV
Time: Thu Nov 21 12:33:25 2019
John -
I had not used the "Display Number of Stats" option before. This
helps. I will try using the boxplot tooI also. I need to think more
about what METViewer is plotting vs what I'm used to seeing from MET.
R/
John
________________________________________
From: John Halley Gotway via RT [met_help at ucar.edu]
Sent: Thursday, November 21, 2019 11:00 AM
To: Raby, John W CIV USARMY CCDC ARL (USA)
Subject: Re: [Non-DoD Source] Re: [rt.rap.ucar.edu #93208] METViewer
v2.11 question
All active links contained in this email were disabled. Please verify
the identity of the sender, and confirm the authenticity of all links
contained within the message prior to copying and pasting the address
to a Web browser.
----
John,
To help clarify what's going on, please go to the "Common" tab and
click
the checkbox for "Display Number of Stats". Then regenerate your plot
and
look at the top axis. For each independent variable value, this tells
you
the number of stats represented by the dot that is plotted for that
lead
time.
When plotting the mean or median of those stats, it makes sense that
the
summary value changes. But neither of those are reported by MET
directly... since Point-Stat is run once for each vx time. If you'd
like
to see the stats computed directly by Point-Stat, I'd recommend
switching
from the "Series" plot type to the "Box" plot type (for boxplots).
John
On Thu, Nov 21, 2019 at 10:28 AM Raby, John W USA CIV via RT <
met_help at ucar.edu> wrote:
>
> <Caution-url: Caution-
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=93208 >
>
> John -
>
> Thanks for this very helpful explanation. In my case, You are
correct in
> your assumption, I ran Point-Stat specifying output line types SL1L2
and
> CNT plus some other line types as well as the .stat output. When I
loaded
> the database, I specified a folder containing the MPR.txt files and
the
> .stat files. See attached load_spec_file. When visualizing in
METViewer, I
> specify STAT for the plot data. I used option (1) for my plotting
which
> generated plots of ME and RMSE as a function of lead time. Based on
your
> info below, the series of ME and RMSE values plotted will come from
the CNT
> line type data in the .stat files. I noticed that when I plot using
the
> "Mean" setting the plot looks different than when using the "Median"
> setting which is expected. Which of these two settings will produce
the ME
> and RMSE values which appear in the CNT.txt file?
>
> By specifying the <load_mpr>true</load_mpr> in the load_spec_file as
> "true", does that load the MPR line type data from the .stat file or
the
> MPR.txt files? I guess my question is what does this line in the
> load_spec_file do? As you will see in the attached load_spec_file, I
also
> specify <load_stat>true</load_stat>, which I assume loads the
database with
> the data from the .stat file. I don't understand why there would be
the
> option of specifying the loading of the MPR files separately when
the MPR
> line type data is present in the .stat file.
>
> R/
> John
>
>
> ________________________________________
> From: John Halley Gotway via RT [met_help at ucar.edu]
> Sent: Thursday, November 21, 2019 9:38 AM
> To: Raby, John W CIV USARMY CCDC ARL (USA)
> Subject: [Non-DoD Source] Re: [rt.rap.ucar.edu #93208] METViewer
v2.11
> question
>
> All active links contained in this email were disabled. Please
verify the
> identity of the sender, and confirm the authenticity of all links
contained
> within the message prior to copying and pasting the address to a Web
> browser.
>
>
>
>
> ----
>
> John,
>
> I see that you have a question about how the stats computed by
METviewer
> compare to those computed by MET.
>
> Let's assume that when you ran Point-Stat, you requested the output
line
> types SL1L2 and CNT. When you load those .stat files into
METviewer, you
> can select specific line types to load, or you can load all of them.
I'll
> assume that you loaded all of them into METviewer.
>
> METviewer supports 3 options in the "Statistics" panel. Below is a
> description of how the MET output is used for each of these options:
>
> (1) Summary: Pull the statistics (e.g. ME or RMSE) from the MET CNT
line
> type directly from the database. For each independent variable
value (i.e.
> value on the x-axis), plot either the "Median", "Mean", or "Sum" of
those
> statistics based on how the "Plot statistic" dropdown option is set.
>
> (2) Aggregation Statistics: When processing continuous statistics,
the
> dropdown list below should be set to "Scalar partial sums (SL1L2)"
(or
> SAL1L2 for anomaly correlation). Rather than selecting the CNT
statistics
> pre-computed by MET, select the partial sums from the SL1L2 line
type. For
> each independent variable value (i.e. value on the x-axis),
aggregate the
> corresponding SL1L2 partial sums together by computing a weighted
average
> where the weight is the number of matched pairs (i.e. TOTAL column)
for
> each line. From those aggregate partial sums, compute and plot a
single
> statistic (e.g. ME or RMSE).
>
> (3) Revision Statistics: This is relatively new. I'm not sure
whether we
> get the statistics using method (1) or (2) above. But once we have
them,
> we sort them in time and compute "revisions", meaning the change in
> statistic value from one time to the next. And it's those revisions
which
> are plotted.
>
> To put is simply, method (1) plots the mean of daily statistics
while
> method (2) plots a single aggregated statistic. Method (1) is
plotting the
> statistics reported directly by Point-Stat.
>
> Thanks,
> John
>
> On Thu, Nov 21, 2019 at 8:10 AM Raby, John W USA CIV via RT <
> met_help at ucar.edu> wrote:
>
> >
> > Thu Nov 21 08:09:37 2019: Request 93208 was acted upon.
> > Transaction: Ticket created by john.w.raby2.civ at mail.mil
> > Queue: met_help
> > Subject: METViewer v2.11 question
> > Owner: Nobody
> > Requestors: john.w.raby2.civ at mail.mil
> > Status: new
> > Ticket <Caution-Caution-url: Caution-
> Caution-https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=93208 >
> >
> >
> > When plotting continuous statistics ME and RMSE using MET input
.stat
> > files from Point-Stat, which settings in the "Statistics" control
panel
> of
> > the web GUI would generate statistics which match those produced
in the
> MET
> > text output file CNT.txt?
> >
> > Thanks.
> >
> > R/
> > John
> >
> > Mr John W. Raby, Meteorologist
> >
> > U.S. Army Research Laboratory
> >
> > White Sands Missile Range, NM 88002
> >
> > (575) 678-2004 DSN 258-2004
> >
> > FAX (575) 678-1230 DSN 258-1230
> >
> > Email: john.w.raby2.civ at mail.mil<Caution-Caution-
mailto:john.raby at us.army.mil>
> >
> >
>
>
>
------------------------------------------------
Subject: RE: [Non-DoD Source] Re: [rt.rap.ucar.edu #93208] METViewer v2.11 question
From: Raby, John W USA CIV
Time: Thu Nov 21 13:11:28 2019
John -
I used the number of stats shown at the top of the plots and compared
that to the number of occurrences of the CNT line type in the .stat
file at a fixed lead time and found that they are equal. Now I am
understanding that the plot was showing the median value of the ME
over all vertical levels at that lead time. The ME values in the CNT
text file are averaged over all matched pairs at a given pressure
level or Z (2 or 10) level. I also produced a box plot which has
expanded my capabilties too. Thanks for your help.
R/
John
________________________________________
From: John Halley Gotway via RT [met_help at ucar.edu]
Sent: Thursday, November 21, 2019 11:00 AM
To: Raby, John W CIV USARMY CCDC ARL (USA)
Subject: Re: [Non-DoD Source] Re: [rt.rap.ucar.edu #93208] METViewer
v2.11 question
All active links contained in this email were disabled. Please verify
the identity of the sender, and confirm the authenticity of all links
contained within the message prior to copying and pasting the address
to a Web browser.
----
John,
To help clarify what's going on, please go to the "Common" tab and
click
the checkbox for "Display Number of Stats". Then regenerate your plot
and
look at the top axis. For each independent variable value, this tells
you
the number of stats represented by the dot that is plotted for that
lead
time.
When plotting the mean or median of those stats, it makes sense that
the
summary value changes. But neither of those are reported by MET
directly... since Point-Stat is run once for each vx time. If you'd
like
to see the stats computed directly by Point-Stat, I'd recommend
switching
from the "Series" plot type to the "Box" plot type (for boxplots).
John
On Thu, Nov 21, 2019 at 10:28 AM Raby, John W USA CIV via RT <
met_help at ucar.edu> wrote:
>
> <Caution-url: Caution-
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=93208 >
>
> John -
>
> Thanks for this very helpful explanation. In my case, You are
correct in
> your assumption, I ran Point-Stat specifying output line types SL1L2
and
> CNT plus some other line types as well as the .stat output. When I
loaded
> the database, I specified a folder containing the MPR.txt files and
the
> .stat files. See attached load_spec_file. When visualizing in
METViewer, I
> specify STAT for the plot data. I used option (1) for my plotting
which
> generated plots of ME and RMSE as a function of lead time. Based on
your
> info below, the series of ME and RMSE values plotted will come from
the CNT
> line type data in the .stat files. I noticed that when I plot using
the
> "Mean" setting the plot looks different than when using the "Median"
> setting which is expected. Which of these two settings will produce
the ME
> and RMSE values which appear in the CNT.txt file?
>
> By specifying the <load_mpr>true</load_mpr> in the load_spec_file as
> "true", does that load the MPR line type data from the .stat file or
the
> MPR.txt files? I guess my question is what does this line in the
> load_spec_file do? As you will see in the attached load_spec_file, I
also
> specify <load_stat>true</load_stat>, which I assume loads the
database with
> the data from the .stat file. I don't understand why there would be
the
> option of specifying the loading of the MPR files separately when
the MPR
> line type data is present in the .stat file.
>
> R/
> John
>
>
> ________________________________________
> From: John Halley Gotway via RT [met_help at ucar.edu]
> Sent: Thursday, November 21, 2019 9:38 AM
> To: Raby, John W CIV USARMY CCDC ARL (USA)
> Subject: [Non-DoD Source] Re: [rt.rap.ucar.edu #93208] METViewer
v2.11
> question
>
> All active links contained in this email were disabled. Please
verify the
> identity of the sender, and confirm the authenticity of all links
contained
> within the message prior to copying and pasting the address to a Web
> browser.
>
>
>
>
> ----
>
> John,
>
> I see that you have a question about how the stats computed by
METviewer
> compare to those computed by MET.
>
> Let's assume that when you ran Point-Stat, you requested the output
line
> types SL1L2 and CNT. When you load those .stat files into
METviewer, you
> can select specific line types to load, or you can load all of them.
I'll
> assume that you loaded all of them into METviewer.
>
> METviewer supports 3 options in the "Statistics" panel. Below is a
> description of how the MET output is used for each of these options:
>
> (1) Summary: Pull the statistics (e.g. ME or RMSE) from the MET CNT
line
> type directly from the database. For each independent variable
value (i.e.
> value on the x-axis), plot either the "Median", "Mean", or "Sum" of
those
> statistics based on how the "Plot statistic" dropdown option is set.
>
> (2) Aggregation Statistics: When processing continuous statistics,
the
> dropdown list below should be set to "Scalar partial sums (SL1L2)"
(or
> SAL1L2 for anomaly correlation). Rather than selecting the CNT
statistics
> pre-computed by MET, select the partial sums from the SL1L2 line
type. For
> each independent variable value (i.e. value on the x-axis),
aggregate the
> corresponding SL1L2 partial sums together by computing a weighted
average
> where the weight is the number of matched pairs (i.e. TOTAL column)
for
> each line. From those aggregate partial sums, compute and plot a
single
> statistic (e.g. ME or RMSE).
>
> (3) Revision Statistics: This is relatively new. I'm not sure
whether we
> get the statistics using method (1) or (2) above. But once we have
them,
> we sort them in time and compute "revisions", meaning the change in
> statistic value from one time to the next. And it's those revisions
which
> are plotted.
>
> To put is simply, method (1) plots the mean of daily statistics
while
> method (2) plots a single aggregated statistic. Method (1) is
plotting the
> statistics reported directly by Point-Stat.
>
> Thanks,
> John
>
> On Thu, Nov 21, 2019 at 8:10 AM Raby, John W USA CIV via RT <
> met_help at ucar.edu> wrote:
>
> >
> > Thu Nov 21 08:09:37 2019: Request 93208 was acted upon.
> > Transaction: Ticket created by john.w.raby2.civ at mail.mil
> > Queue: met_help
> > Subject: METViewer v2.11 question
> > Owner: Nobody
> > Requestors: john.w.raby2.civ at mail.mil
> > Status: new
> > Ticket <Caution-Caution-url: Caution-
> Caution-https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=93208 >
> >
> >
> > When plotting continuous statistics ME and RMSE using MET input
.stat
> > files from Point-Stat, which settings in the "Statistics" control
panel
> of
> > the web GUI would generate statistics which match those produced
in the
> MET
> > text output file CNT.txt?
> >
> > Thanks.
> >
> > R/
> > John
> >
> > Mr John W. Raby, Meteorologist
> >
> > U.S. Army Research Laboratory
> >
> > White Sands Missile Range, NM 88002
> >
> > (575) 678-2004 DSN 258-2004
> >
> > FAX (575) 678-1230 DSN 258-1230
> >
> > Email: john.w.raby2.civ at mail.mil<Caution-Caution-
mailto:john.raby at us.army.mil>
> >
> >
>
>
>
------------------------------------------------
Subject: METViewer v2.11 question
From: John Halley Gotway
Time: Thu Nov 21 13:26:46 2019
John,
And I failed to answer your other question about the MPR line type.
Yes,
it can be loaded in METviewer. And the values from that line type can
be
plotted (e.g. MPR_FCST, MPR_OBS). However, METviewer does currently
derive
any stats, like ME or RMSE, from the MPR line type.
It is very important to realize that METviewer will silently combine
all
results that meet your search criteria. If you don't specify a level,
it'll use all levels available in the database. When creating a new
plot,
I recommend checking all of the values for the header columns in the
"Fixed
Values" section (e.g. VX_MASK, INTERP_MTHD, INTERP_PNTS, FCST_LEV, and
so
on). Make sure that you're plotting the subset of data that you
intend.
And turn on and check the number of stats as a sanity check. You can
also
click on the "R data" tab to see the actual data values included in
that
plot. That's another sanity check.
Thanks,
John
On Thu, Nov 21, 2019 at 12:33 PM Raby, John W USA CIV via RT <
met_help at ucar.edu> wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=93208 >
>
> John -
>
> I had not used the "Display Number of Stats" option before. This
helps. I
> will try using the boxplot tooI also. I need to think more about
what
> METViewer is plotting vs what I'm used to seeing from MET.
> R/
> John
>
> ________________________________________
> From: John Halley Gotway via RT [met_help at ucar.edu]
> Sent: Thursday, November 21, 2019 11:00 AM
> To: Raby, John W CIV USARMY CCDC ARL (USA)
> Subject: Re: [Non-DoD Source] Re: [rt.rap.ucar.edu #93208] METViewer
> v2.11 question
>
> All active links contained in this email were disabled. Please
verify the
> identity of the sender, and confirm the authenticity of all links
contained
> within the message prior to copying and pasting the address to a Web
> browser.
>
>
>
>
> ----
>
> John,
>
> To help clarify what's going on, please go to the "Common" tab and
click
> the checkbox for "Display Number of Stats". Then regenerate your
plot and
> look at the top axis. For each independent variable value, this
tells you
> the number of stats represented by the dot that is plotted for that
lead
> time.
>
> When plotting the mean or median of those stats, it makes sense that
the
> summary value changes. But neither of those are reported by MET
> directly... since Point-Stat is run once for each vx time. If you'd
like
> to see the stats computed directly by Point-Stat, I'd recommend
switching
> from the "Series" plot type to the "Box" plot type (for boxplots).
>
> John
>
> On Thu, Nov 21, 2019 at 10:28 AM Raby, John W USA CIV via RT <
> met_help at ucar.edu> wrote:
>
> >
> > <Caution-url: Caution-
> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=93208 >
> >
> > John -
> >
> > Thanks for this very helpful explanation. In my case, You are
correct in
> > your assumption, I ran Point-Stat specifying output line types
SL1L2 and
> > CNT plus some other line types as well as the .stat output. When I
loaded
> > the database, I specified a folder containing the MPR.txt files
and the
> > .stat files. See attached load_spec_file. When visualizing in
METViewer,
> I
> > specify STAT for the plot data. I used option (1) for my plotting
which
> > generated plots of ME and RMSE as a function of lead time. Based
on your
> > info below, the series of ME and RMSE values plotted will come
from the
> CNT
> > line type data in the .stat files. I noticed that when I plot
using the
> > "Mean" setting the plot looks different than when using the
"Median"
> > setting which is expected. Which of these two settings will
produce the
> ME
> > and RMSE values which appear in the CNT.txt file?
> >
> > By specifying the <load_mpr>true</load_mpr> in the load_spec_file
as
> > "true", does that load the MPR line type data from the .stat file
or the
> > MPR.txt files? I guess my question is what does this line in the
> > load_spec_file do? As you will see in the attached load_spec_file,
I also
> > specify <load_stat>true</load_stat>, which I assume loads the
database
> with
> > the data from the .stat file. I don't understand why there would
be the
> > option of specifying the loading of the MPR files separately when
the MPR
> > line type data is present in the .stat file.
> >
> > R/
> > John
> >
> >
> > ________________________________________
> > From: John Halley Gotway via RT [met_help at ucar.edu]
> > Sent: Thursday, November 21, 2019 9:38 AM
> > To: Raby, John W CIV USARMY CCDC ARL (USA)
> > Subject: [Non-DoD Source] Re: [rt.rap.ucar.edu #93208] METViewer
v2.11
> > question
> >
> > All active links contained in this email were disabled. Please
verify
> the
> > identity of the sender, and confirm the authenticity of all links
> contained
> > within the message prior to copying and pasting the address to a
Web
> > browser.
> >
> >
> >
> >
> > ----
> >
> > John,
> >
> > I see that you have a question about how the stats computed by
METviewer
> > compare to those computed by MET.
> >
> > Let's assume that when you ran Point-Stat, you requested the
output line
> > types SL1L2 and CNT. When you load those .stat files into
METviewer, you
> > can select specific line types to load, or you can load all of
them.
> I'll
> > assume that you loaded all of them into METviewer.
> >
> > METviewer supports 3 options in the "Statistics" panel. Below is
a
> > description of how the MET output is used for each of these
options:
> >
> > (1) Summary: Pull the statistics (e.g. ME or RMSE) from the MET
CNT line
> > type directly from the database. For each independent variable
value
> (i.e.
> > value on the x-axis), plot either the "Median", "Mean", or "Sum"
of those
> > statistics based on how the "Plot statistic" dropdown option is
set.
> >
> > (2) Aggregation Statistics: When processing continuous statistics,
the
> > dropdown list below should be set to "Scalar partial sums (SL1L2)"
(or
> > SAL1L2 for anomaly correlation). Rather than selecting the CNT
> statistics
> > pre-computed by MET, select the partial sums from the SL1L2 line
type.
> For
> > each independent variable value (i.e. value on the x-axis),
aggregate the
> > corresponding SL1L2 partial sums together by computing a weighted
average
> > where the weight is the number of matched pairs (i.e. TOTAL
column) for
> > each line. From those aggregate partial sums, compute and plot a
single
> > statistic (e.g. ME or RMSE).
> >
> > (3) Revision Statistics: This is relatively new. I'm not sure
whether we
> > get the statistics using method (1) or (2) above. But once we
have them,
> > we sort them in time and compute "revisions", meaning the change
in
> > statistic value from one time to the next. And it's those
revisions
> which
> > are plotted.
> >
> > To put is simply, method (1) plots the mean of daily statistics
while
> > method (2) plots a single aggregated statistic. Method (1) is
plotting
> the
> > statistics reported directly by Point-Stat.
> >
> > Thanks,
> > John
> >
> > On Thu, Nov 21, 2019 at 8:10 AM Raby, John W USA CIV via RT <
> > met_help at ucar.edu> wrote:
> >
> > >
> > > Thu Nov 21 08:09:37 2019: Request 93208 was acted upon.
> > > Transaction: Ticket created by john.w.raby2.civ at mail.mil
> > > Queue: met_help
> > > Subject: METViewer v2.11 question
> > > Owner: Nobody
> > > Requestors: john.w.raby2.civ at mail.mil
> > > Status: new
> > > Ticket <Caution-Caution-url: Caution-
> > Caution-https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=93208 >
> > >
> > >
> > > When plotting continuous statistics ME and RMSE using MET input
.stat
> > > files from Point-Stat, which settings in the "Statistics"
control panel
> > of
> > > the web GUI would generate statistics which match those produced
in the
> > MET
> > > text output file CNT.txt?
> > >
> > > Thanks.
> > >
> > > R/
> > > John
> > >
> > > Mr John W. Raby, Meteorologist
> > >
> > > U.S. Army Research Laboratory
> > >
> > > White Sands Missile Range, NM 88002
> > >
> > > (575) 678-2004 DSN 258-2004
> > >
> > > FAX (575) 678-1230 DSN 258-1230
> > >
> > > Email: john.w.raby2.civ at mail.mil<Caution-Caution-mailto:
> john.raby at us.army.mil>
> > >
> > >
> >
> >
> >
>
>
>
>
------------------------------------------------
Subject: RE: [Non-DoD Source] Re: [rt.rap.ucar.edu #93208] METViewer v2.11 question
From: Raby, John W USA CIV
Time: Thu Nov 21 13:41:53 2019
John -
Thanks for sharing this good info on now METViewer works. This has
really helped. I just used the "Fixed Value" option to restrict the ME
values being used for the plot I was working with and now I am
displaying only the ME values at the Z2 level which is matching the
values from the CNT text file exactly.
For the MPR line type, is it reading the line type data from the .stat
file or form the MPR.txt file? I just plotted MPR_OBS for the Z2 level
and it appears to be plotting the median value of all observations at
that level.
R/
John
________________________________________
From: John Halley Gotway via RT [met_help at ucar.edu]
Sent: Thursday, November 21, 2019 1:26 PM
To: Raby, John W CIV USARMY CCDC ARL (USA)
Subject: Re: [Non-DoD Source] Re: [rt.rap.ucar.edu #93208] METViewer
v2.11 question
All active links contained in this email were disabled. Please verify
the identity of the sender, and confirm the authenticity of all links
contained within the message prior to copying and pasting the address
to a Web browser.
----
John,
And I failed to answer your other question about the MPR line type.
Yes,
it can be loaded in METviewer. And the values from that line type can
be
plotted (e.g. MPR_FCST, MPR_OBS). However, METviewer does currently
derive
any stats, like ME or RMSE, from the MPR line type.
It is very important to realize that METviewer will silently combine
all
results that meet your search criteria. If you don't specify a level,
it'll use all levels available in the database. When creating a new
plot,
I recommend checking all of the values for the header columns in the
"Fixed
Values" section (e.g. VX_MASK, INTERP_MTHD, INTERP_PNTS, FCST_LEV, and
so
on). Make sure that you're plotting the subset of data that you
intend.
And turn on and check the number of stats as a sanity check. You can
also
click on the "R data" tab to see the actual data values included in
that
plot. That's another sanity check.
Thanks,
John
On Thu, Nov 21, 2019 at 12:33 PM Raby, John W USA CIV via RT <
met_help at ucar.edu> wrote:
>
> <Caution-url: Caution-
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=93208 >
>
> John -
>
> I had not used the "Display Number of Stats" option before. This
helps. I
> will try using the boxplot tooI also. I need to think more about
what
> METViewer is plotting vs what I'm used to seeing from MET.
> R/
> John
>
> ________________________________________
> From: John Halley Gotway via RT [met_help at ucar.edu]
> Sent: Thursday, November 21, 2019 11:00 AM
> To: Raby, John W CIV USARMY CCDC ARL (USA)
> Subject: Re: [Non-DoD Source] Re: [rt.rap.ucar.edu #93208] METViewer
> v2.11 question
>
> All active links contained in this email were disabled. Please
verify the
> identity of the sender, and confirm the authenticity of all links
contained
> within the message prior to copying and pasting the address to a Web
> browser.
>
>
>
>
> ----
>
> John,
>
> To help clarify what's going on, please go to the "Common" tab and
click
> the checkbox for "Display Number of Stats". Then regenerate your
plot and
> look at the top axis. For each independent variable value, this
tells you
> the number of stats represented by the dot that is plotted for that
lead
> time.
>
> When plotting the mean or median of those stats, it makes sense that
the
> summary value changes. But neither of those are reported by MET
> directly... since Point-Stat is run once for each vx time. If you'd
like
> to see the stats computed directly by Point-Stat, I'd recommend
switching
> from the "Series" plot type to the "Box" plot type (for boxplots).
>
> John
>
> On Thu, Nov 21, 2019 at 10:28 AM Raby, John W USA CIV via RT <
> met_help at ucar.edu> wrote:
>
> >
> > <Caution-Caution-url: Caution-
> Caution-https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=93208 >
> >
> > John -
> >
> > Thanks for this very helpful explanation. In my case, You are
correct in
> > your assumption, I ran Point-Stat specifying output line types
SL1L2 and
> > CNT plus some other line types as well as the .stat output. When I
loaded
> > the database, I specified a folder containing the MPR.txt files
and the
> > .stat files. See attached load_spec_file. When visualizing in
METViewer,
> I
> > specify STAT for the plot data. I used option (1) for my plotting
which
> > generated plots of ME and RMSE as a function of lead time. Based
on your
> > info below, the series of ME and RMSE values plotted will come
from the
> CNT
> > line type data in the .stat files. I noticed that when I plot
using the
> > "Mean" setting the plot looks different than when using the
"Median"
> > setting which is expected. Which of these two settings will
produce the
> ME
> > and RMSE values which appear in the CNT.txt file?
> >
> > By specifying the <load_mpr>true</load_mpr> in the load_spec_file
as
> > "true", does that load the MPR line type data from the .stat file
or the
> > MPR.txt files? I guess my question is what does this line in the
> > load_spec_file do? As you will see in the attached load_spec_file,
I also
> > specify <load_stat>true</load_stat>, which I assume loads the
database
> with
> > the data from the .stat file. I don't understand why there would
be the
> > option of specifying the loading of the MPR files separately when
the MPR
> > line type data is present in the .stat file.
> >
> > R/
> > John
> >
> >
> > ________________________________________
> > From: John Halley Gotway via RT [met_help at ucar.edu]
> > Sent: Thursday, November 21, 2019 9:38 AM
> > To: Raby, John W CIV USARMY CCDC ARL (USA)
> > Subject: [Non-DoD Source] Re: [rt.rap.ucar.edu #93208] METViewer
v2.11
> > question
> >
> > All active links contained in this email were disabled. Please
verify
> the
> > identity of the sender, and confirm the authenticity of all links
> contained
> > within the message prior to copying and pasting the address to a
Web
> > browser.
> >
> >
> >
> >
> > ----
> >
> > John,
> >
> > I see that you have a question about how the stats computed by
METviewer
> > compare to those computed by MET.
> >
> > Let's assume that when you ran Point-Stat, you requested the
output line
> > types SL1L2 and CNT. When you load those .stat files into
METviewer, you
> > can select specific line types to load, or you can load all of
them.
> I'll
> > assume that you loaded all of them into METviewer.
> >
> > METviewer supports 3 options in the "Statistics" panel. Below is
a
> > description of how the MET output is used for each of these
options:
> >
> > (1) Summary: Pull the statistics (e.g. ME or RMSE) from the MET
CNT line
> > type directly from the database. For each independent variable
value
> (i.e.
> > value on the x-axis), plot either the "Median", "Mean", or "Sum"
of those
> > statistics based on how the "Plot statistic" dropdown option is
set.
> >
> > (2) Aggregation Statistics: When processing continuous statistics,
the
> > dropdown list below should be set to "Scalar partial sums (SL1L2)"
(or
> > SAL1L2 for anomaly correlation). Rather than selecting the CNT
> statistics
> > pre-computed by MET, select the partial sums from the SL1L2 line
type.
> For
> > each independent variable value (i.e. value on the x-axis),
aggregate the
> > corresponding SL1L2 partial sums together by computing a weighted
average
> > where the weight is the number of matched pairs (i.e. TOTAL
column) for
> > each line. From those aggregate partial sums, compute and plot a
single
> > statistic (e.g. ME or RMSE).
> >
> > (3) Revision Statistics: This is relatively new. I'm not sure
whether we
> > get the statistics using method (1) or (2) above. But once we
have them,
> > we sort them in time and compute "revisions", meaning the change
in
> > statistic value from one time to the next. And it's those
revisions
> which
> > are plotted.
> >
> > To put is simply, method (1) plots the mean of daily statistics
while
> > method (2) plots a single aggregated statistic. Method (1) is
plotting
> the
> > statistics reported directly by Point-Stat.
> >
> > Thanks,
> > John
> >
> > On Thu, Nov 21, 2019 at 8:10 AM Raby, John W USA CIV via RT <
> > met_help at ucar.edu> wrote:
> >
> > >
> > > Thu Nov 21 08:09:37 2019: Request 93208 was acted upon.
> > > Transaction: Ticket created by john.w.raby2.civ at mail.mil
> > > Queue: met_help
> > > Subject: METViewer v2.11 question
> > > Owner: Nobody
> > > Requestors: john.w.raby2.civ at mail.mil
> > > Status: new
> > > Ticket <Caution-Caution-Caution-url: Caution-
> > Caution-Caution-
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=93208 >
> > >
> > >
> > > When plotting continuous statistics ME and RMSE using MET input
.stat
> > > files from Point-Stat, which settings in the "Statistics"
control panel
> > of
> > > the web GUI would generate statistics which match those produced
in the
> > MET
> > > text output file CNT.txt?
> > >
> > > Thanks.
> > >
> > > R/
> > > John
> > >
> > > Mr John W. Raby, Meteorologist
> > >
> > > U.S. Army Research Laboratory
> > >
> > > White Sands Missile Range, NM 88002
> > >
> > > (575) 678-2004 DSN 258-2004
> > >
> > > FAX (575) 678-1230 DSN 258-1230
> > >
> > > Email: john.w.raby2.civ at mail.mil<Caution-Caution-Caution-mailto:
> john.raby at us.army.mil>
> > >
> > >
> >
> >
> >
>
>
>
>
------------------------------------------------
Subject: METViewer v2.11 question
From: John Halley Gotway
Time: Thu Nov 21 13:59:19 2019
John,
When you run the load script, METviewer is *only* loading data from
the
.stat output files, never the .txt files. If/when you get to the
point of
exclusively using METviewer to interact with the output from MET, then
you
could consider turning off the writing of those options text files.
The
contents of the text files are just a duplication of the contents of
the
.stat file. They are just easier for humans (and post-processing
scripts)
to read.
John
On Thu, Nov 21, 2019 at 1:42 PM Raby, John W USA CIV via RT <
met_help at ucar.edu> wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=93208 >
>
> John -
>
> Thanks for sharing this good info on now METViewer works. This has
really
> helped. I just used the "Fixed Value" option to restrict the ME
values
> being used for the plot I was working with and now I am displaying
only the
> ME values at the Z2 level which is matching the values from the CNT
text
> file exactly.
>
> For the MPR line type, is it reading the line type data from the
.stat
> file or form the MPR.txt file? I just plotted MPR_OBS for the Z2
level and
> it appears to be plotting the median value of all observations at
that
> level.
>
> R/
> John
> ________________________________________
> From: John Halley Gotway via RT [met_help at ucar.edu]
> Sent: Thursday, November 21, 2019 1:26 PM
> To: Raby, John W CIV USARMY CCDC ARL (USA)
> Subject: Re: [Non-DoD Source] Re: [rt.rap.ucar.edu #93208] METViewer
> v2.11 question
>
> All active links contained in this email were disabled. Please
verify the
> identity of the sender, and confirm the authenticity of all links
contained
> within the message prior to copying and pasting the address to a Web
> browser.
>
>
>
>
> ----
>
> John,
>
> And I failed to answer your other question about the MPR line type.
Yes,
> it can be loaded in METviewer. And the values from that line type
can be
> plotted (e.g. MPR_FCST, MPR_OBS). However, METviewer does currently
derive
> any stats, like ME or RMSE, from the MPR line type.
>
> It is very important to realize that METviewer will silently combine
all
> results that meet your search criteria. If you don't specify a
level,
> it'll use all levels available in the database. When creating a new
plot,
> I recommend checking all of the values for the header columns in the
"Fixed
> Values" section (e.g. VX_MASK, INTERP_MTHD, INTERP_PNTS, FCST_LEV,
and so
> on). Make sure that you're plotting the subset of data that you
intend.
> And turn on and check the number of stats as a sanity check. You
can also
> click on the "R data" tab to see the actual data values included in
that
> plot. That's another sanity check.
>
> Thanks,
> John
>
> On Thu, Nov 21, 2019 at 12:33 PM Raby, John W USA CIV via RT <
> met_help at ucar.edu> wrote:
>
> >
> > <Caution-url: Caution-
> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=93208 >
> >
> > John -
> >
> > I had not used the "Display Number of Stats" option before. This
helps. I
> > will try using the boxplot tooI also. I need to think more about
what
> > METViewer is plotting vs what I'm used to seeing from MET.
> > R/
> > John
> >
> > ________________________________________
> > From: John Halley Gotway via RT [met_help at ucar.edu]
> > Sent: Thursday, November 21, 2019 11:00 AM
> > To: Raby, John W CIV USARMY CCDC ARL (USA)
> > Subject: Re: [Non-DoD Source] Re: [rt.rap.ucar.edu #93208]
METViewer
> > v2.11 question
> >
> > All active links contained in this email were disabled. Please
verify
> the
> > identity of the sender, and confirm the authenticity of all links
> contained
> > within the message prior to copying and pasting the address to a
Web
> > browser.
> >
> >
> >
> >
> > ----
> >
> > John,
> >
> > To help clarify what's going on, please go to the "Common" tab and
click
> > the checkbox for "Display Number of Stats". Then regenerate your
plot
> and
> > look at the top axis. For each independent variable value, this
tells
> you
> > the number of stats represented by the dot that is plotted for
that lead
> > time.
> >
> > When plotting the mean or median of those stats, it makes sense
that the
> > summary value changes. But neither of those are reported by MET
> > directly... since Point-Stat is run once for each vx time. If
you'd like
> > to see the stats computed directly by Point-Stat, I'd recommend
switching
> > from the "Series" plot type to the "Box" plot type (for boxplots).
> >
> > John
> >
> > On Thu, Nov 21, 2019 at 10:28 AM Raby, John W USA CIV via RT <
> > met_help at ucar.edu> wrote:
> >
> > >
> > > <Caution-Caution-url: Caution-
> > Caution-https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=93208 >
> > >
> > > John -
> > >
> > > Thanks for this very helpful explanation. In my case, You are
correct
> in
> > > your assumption, I ran Point-Stat specifying output line types
SL1L2
> and
> > > CNT plus some other line types as well as the .stat output. When
I
> loaded
> > > the database, I specified a folder containing the MPR.txt files
and the
> > > .stat files. See attached load_spec_file. When visualizing in
> METViewer,
> > I
> > > specify STAT for the plot data. I used option (1) for my
plotting which
> > > generated plots of ME and RMSE as a function of lead time. Based
on
> your
> > > info below, the series of ME and RMSE values plotted will come
from the
> > CNT
> > > line type data in the .stat files. I noticed that when I plot
using the
> > > "Mean" setting the plot looks different than when using the
"Median"
> > > setting which is expected. Which of these two settings will
produce the
> > ME
> > > and RMSE values which appear in the CNT.txt file?
> > >
> > > By specifying the <load_mpr>true</load_mpr> in the
load_spec_file as
> > > "true", does that load the MPR line type data from the .stat
file or
> the
> > > MPR.txt files? I guess my question is what does this line in the
> > > load_spec_file do? As you will see in the attached
load_spec_file, I
> also
> > > specify <load_stat>true</load_stat>, which I assume loads the
database
> > with
> > > the data from the .stat file. I don't understand why there would
be the
> > > option of specifying the loading of the MPR files separately
when the
> MPR
> > > line type data is present in the .stat file.
> > >
> > > R/
> > > John
> > >
> > >
> > > ________________________________________
> > > From: John Halley Gotway via RT [met_help at ucar.edu]
> > > Sent: Thursday, November 21, 2019 9:38 AM
> > > To: Raby, John W CIV USARMY CCDC ARL (USA)
> > > Subject: [Non-DoD Source] Re: [rt.rap.ucar.edu #93208] METViewer
v2.11
> > > question
> > >
> > > All active links contained in this email were disabled. Please
verify
> > the
> > > identity of the sender, and confirm the authenticity of all
links
> > contained
> > > within the message prior to copying and pasting the address to a
Web
> > > browser.
> > >
> > >
> > >
> > >
> > > ----
> > >
> > > John,
> > >
> > > I see that you have a question about how the stats computed by
> METviewer
> > > compare to those computed by MET.
> > >
> > > Let's assume that when you ran Point-Stat, you requested the
output
> line
> > > types SL1L2 and CNT. When you load those .stat files into
METviewer,
> you
> > > can select specific line types to load, or you can load all of
them.
> > I'll
> > > assume that you loaded all of them into METviewer.
> > >
> > > METviewer supports 3 options in the "Statistics" panel. Below
is a
> > > description of how the MET output is used for each of these
options:
> > >
> > > (1) Summary: Pull the statistics (e.g. ME or RMSE) from the MET
CNT
> line
> > > type directly from the database. For each independent variable
value
> > (i.e.
> > > value on the x-axis), plot either the "Median", "Mean", or "Sum"
of
> those
> > > statistics based on how the "Plot statistic" dropdown option is
set.
> > >
> > > (2) Aggregation Statistics: When processing continuous
statistics, the
> > > dropdown list below should be set to "Scalar partial sums
(SL1L2)" (or
> > > SAL1L2 for anomaly correlation). Rather than selecting the CNT
> > statistics
> > > pre-computed by MET, select the partial sums from the SL1L2 line
type.
> > For
> > > each independent variable value (i.e. value on the x-axis),
aggregate
> the
> > > corresponding SL1L2 partial sums together by computing a
weighted
> average
> > > where the weight is the number of matched pairs (i.e. TOTAL
column) for
> > > each line. From those aggregate partial sums, compute and plot
a
> single
> > > statistic (e.g. ME or RMSE).
> > >
> > > (3) Revision Statistics: This is relatively new. I'm not sure
whether
> we
> > > get the statistics using method (1) or (2) above. But once we
have
> them,
> > > we sort them in time and compute "revisions", meaning the change
in
> > > statistic value from one time to the next. And it's those
revisions
> > which
> > > are plotted.
> > >
> > > To put is simply, method (1) plots the mean of daily statistics
while
> > > method (2) plots a single aggregated statistic. Method (1) is
plotting
> > the
> > > statistics reported directly by Point-Stat.
> > >
> > > Thanks,
> > > John
> > >
> > > On Thu, Nov 21, 2019 at 8:10 AM Raby, John W USA CIV via RT <
> > > met_help at ucar.edu> wrote:
> > >
> > > >
> > > > Thu Nov 21 08:09:37 2019: Request 93208 was acted upon.
> > > > Transaction: Ticket created by john.w.raby2.civ at mail.mil
> > > > Queue: met_help
> > > > Subject: METViewer v2.11 question
> > > > Owner: Nobody
> > > > Requestors: john.w.raby2.civ at mail.mil
> > > > Status: new
> > > > Ticket <Caution-Caution-Caution-url: Caution-
> > > Caution-Caution-
> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=93208 >
> > > >
> > > >
> > > > When plotting continuous statistics ME and RMSE using MET
input .stat
> > > > files from Point-Stat, which settings in the "Statistics"
control
> panel
> > > of
> > > > the web GUI would generate statistics which match those
produced in
> the
> > > MET
> > > > text output file CNT.txt?
> > > >
> > > > Thanks.
> > > >
> > > > R/
> > > > John
> > > >
> > > > Mr John W. Raby, Meteorologist
> > > >
> > > > U.S. Army Research Laboratory
> > > >
> > > > White Sands Missile Range, NM 88002
> > > >
> > > > (575) 678-2004 DSN 258-2004
> > > >
> > > > FAX (575) 678-1230 DSN 258-1230
> > > >
> > > > Email: john.w.raby2.civ at mail.mil<Caution-Caution-Caution-
mailto:
> > john.raby at us.army.mil>
> > > >
> > > >
> > >
> > >
> > >
> >
> >
> >
> >
>
>
>
>
------------------------------------------------
Subject: RE: [Non-DoD Source] Re: [rt.rap.ucar.edu #93208] METViewer v2.11 question
From: Raby, John W USA CIV
Time: Thu Nov 21 14:04:49 2019
Thanks for clarifying that, John.
R/
John
________________________________________
From: John Halley Gotway via RT [met_help at ucar.edu]
Sent: Thursday, November 21, 2019 1:59 PM
To: Raby, John W CIV USARMY CCDC ARL (USA)
Subject: Re: [Non-DoD Source] Re: [rt.rap.ucar.edu #93208] METViewer
v2.11 question
All active links contained in this email were disabled. Please verify
the identity of the sender, and confirm the authenticity of all links
contained within the message prior to copying and pasting the address
to a Web browser.
----
John,
When you run the load script, METviewer is *only* loading data from
the
.stat output files, never the .txt files. If/when you get to the
point of
exclusively using METviewer to interact with the output from MET, then
you
could consider turning off the writing of those options text files.
The
contents of the text files are just a duplication of the contents of
the
.stat file. They are just easier for humans (and post-processing
scripts)
to read.
John
On Thu, Nov 21, 2019 at 1:42 PM Raby, John W USA CIV via RT <
met_help at ucar.edu> wrote:
>
> <Caution-url: Caution-
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=93208 >
>
> John -
>
> Thanks for sharing this good info on now METViewer works. This has
really
> helped. I just used the "Fixed Value" option to restrict the ME
values
> being used for the plot I was working with and now I am displaying
only the
> ME values at the Z2 level which is matching the values from the CNT
text
> file exactly.
>
> For the MPR line type, is it reading the line type data from the
.stat
> file or form the MPR.txt file? I just plotted MPR_OBS for the Z2
level and
> it appears to be plotting the median value of all observations at
that
> level.
>
> R/
> John
> ________________________________________
> From: John Halley Gotway via RT [met_help at ucar.edu]
> Sent: Thursday, November 21, 2019 1:26 PM
> To: Raby, John W CIV USARMY CCDC ARL (USA)
> Subject: Re: [Non-DoD Source] Re: [rt.rap.ucar.edu #93208] METViewer
> v2.11 question
>
> All active links contained in this email were disabled. Please
verify the
> identity of the sender, and confirm the authenticity of all links
contained
> within the message prior to copying and pasting the address to a Web
> browser.
>
>
>
>
> ----
>
> John,
>
> And I failed to answer your other question about the MPR line type.
Yes,
> it can be loaded in METviewer. And the values from that line type
can be
> plotted (e.g. MPR_FCST, MPR_OBS). However, METviewer does currently
derive
> any stats, like ME or RMSE, from the MPR line type.
>
> It is very important to realize that METviewer will silently combine
all
> results that meet your search criteria. If you don't specify a
level,
> it'll use all levels available in the database. When creating a new
plot,
> I recommend checking all of the values for the header columns in the
"Fixed
> Values" section (e.g. VX_MASK, INTERP_MTHD, INTERP_PNTS, FCST_LEV,
and so
> on). Make sure that you're plotting the subset of data that you
intend.
> And turn on and check the number of stats as a sanity check. You
can also
> click on the "R data" tab to see the actual data values included in
that
> plot. That's another sanity check.
>
> Thanks,
> John
>
> On Thu, Nov 21, 2019 at 12:33 PM Raby, John W USA CIV via RT <
> met_help at ucar.edu> wrote:
>
> >
> > <Caution-Caution-url: Caution-
> Caution-https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=93208 >
> >
> > John -
> >
> > I had not used the "Display Number of Stats" option before. This
helps. I
> > will try using the boxplot tooI also. I need to think more about
what
> > METViewer is plotting vs what I'm used to seeing from MET.
> > R/
> > John
> >
> > ________________________________________
> > From: John Halley Gotway via RT [met_help at ucar.edu]
> > Sent: Thursday, November 21, 2019 11:00 AM
> > To: Raby, John W CIV USARMY CCDC ARL (USA)
> > Subject: Re: [Non-DoD Source] Re: [rt.rap.ucar.edu #93208]
METViewer
> > v2.11 question
> >
> > All active links contained in this email were disabled. Please
verify
> the
> > identity of the sender, and confirm the authenticity of all links
> contained
> > within the message prior to copying and pasting the address to a
Web
> > browser.
> >
> >
> >
> >
> > ----
> >
> > John,
> >
> > To help clarify what's going on, please go to the "Common" tab and
click
> > the checkbox for "Display Number of Stats". Then regenerate your
plot
> and
> > look at the top axis. For each independent variable value, this
tells
> you
> > the number of stats represented by the dot that is plotted for
that lead
> > time.
> >
> > When plotting the mean or median of those stats, it makes sense
that the
> > summary value changes. But neither of those are reported by MET
> > directly... since Point-Stat is run once for each vx time. If
you'd like
> > to see the stats computed directly by Point-Stat, I'd recommend
switching
> > from the "Series" plot type to the "Box" plot type (for boxplots).
> >
> > John
> >
> > On Thu, Nov 21, 2019 at 10:28 AM Raby, John W USA CIV via RT <
> > met_help at ucar.edu> wrote:
> >
> > >
> > > <Caution-Caution-Caution-url: Caution-
> > Caution-Caution-
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=93208 >
> > >
> > > John -
> > >
> > > Thanks for this very helpful explanation. In my case, You are
correct
> in
> > > your assumption, I ran Point-Stat specifying output line types
SL1L2
> and
> > > CNT plus some other line types as well as the .stat output. When
I
> loaded
> > > the database, I specified a folder containing the MPR.txt files
and the
> > > .stat files. See attached load_spec_file. When visualizing in
> METViewer,
> > I
> > > specify STAT for the plot data. I used option (1) for my
plotting which
> > > generated plots of ME and RMSE as a function of lead time. Based
on
> your
> > > info below, the series of ME and RMSE values plotted will come
from the
> > CNT
> > > line type data in the .stat files. I noticed that when I plot
using the
> > > "Mean" setting the plot looks different than when using the
"Median"
> > > setting which is expected. Which of these two settings will
produce the
> > ME
> > > and RMSE values which appear in the CNT.txt file?
> > >
> > > By specifying the <load_mpr>true</load_mpr> in the
load_spec_file as
> > > "true", does that load the MPR line type data from the .stat
file or
> the
> > > MPR.txt files? I guess my question is what does this line in the
> > > load_spec_file do? As you will see in the attached
load_spec_file, I
> also
> > > specify <load_stat>true</load_stat>, which I assume loads the
database
> > with
> > > the data from the .stat file. I don't understand why there would
be the
> > > option of specifying the loading of the MPR files separately
when the
> MPR
> > > line type data is present in the .stat file.
> > >
> > > R/
> > > John
> > >
> > >
> > > ________________________________________
> > > From: John Halley Gotway via RT [met_help at ucar.edu]
> > > Sent: Thursday, November 21, 2019 9:38 AM
> > > To: Raby, John W CIV USARMY CCDC ARL (USA)
> > > Subject: [Non-DoD Source] Re: [rt.rap.ucar.edu #93208] METViewer
v2.11
> > > question
> > >
> > > All active links contained in this email were disabled. Please
verify
> > the
> > > identity of the sender, and confirm the authenticity of all
links
> > contained
> > > within the message prior to copying and pasting the address to a
Web
> > > browser.
> > >
> > >
> > >
> > >
> > > ----
> > >
> > > John,
> > >
> > > I see that you have a question about how the stats computed by
> METviewer
> > > compare to those computed by MET.
> > >
> > > Let's assume that when you ran Point-Stat, you requested the
output
> line
> > > types SL1L2 and CNT. When you load those .stat files into
METviewer,
> you
> > > can select specific line types to load, or you can load all of
them.
> > I'll
> > > assume that you loaded all of them into METviewer.
> > >
> > > METviewer supports 3 options in the "Statistics" panel. Below
is a
> > > description of how the MET output is used for each of these
options:
> > >
> > > (1) Summary: Pull the statistics (e.g. ME or RMSE) from the MET
CNT
> line
> > > type directly from the database. For each independent variable
value
> > (i.e.
> > > value on the x-axis), plot either the "Median", "Mean", or "Sum"
of
> those
> > > statistics based on how the "Plot statistic" dropdown option is
set.
> > >
> > > (2) Aggregation Statistics: When processing continuous
statistics, the
> > > dropdown list below should be set to "Scalar partial sums
(SL1L2)" (or
> > > SAL1L2 for anomaly correlation). Rather than selecting the CNT
> > statistics
> > > pre-computed by MET, select the partial sums from the SL1L2 line
type.
> > For
> > > each independent variable value (i.e. value on the x-axis),
aggregate
> the
> > > corresponding SL1L2 partial sums together by computing a
weighted
> average
> > > where the weight is the number of matched pairs (i.e. TOTAL
column) for
> > > each line. From those aggregate partial sums, compute and plot
a
> single
> > > statistic (e.g. ME or RMSE).
> > >
> > > (3) Revision Statistics: This is relatively new. I'm not sure
whether
> we
> > > get the statistics using method (1) or (2) above. But once we
have
> them,
> > > we sort them in time and compute "revisions", meaning the change
in
> > > statistic value from one time to the next. And it's those
revisions
> > which
> > > are plotted.
> > >
> > > To put is simply, method (1) plots the mean of daily statistics
while
> > > method (2) plots a single aggregated statistic. Method (1) is
plotting
> > the
> > > statistics reported directly by Point-Stat.
> > >
> > > Thanks,
> > > John
> > >
> > > On Thu, Nov 21, 2019 at 8:10 AM Raby, John W USA CIV via RT <
> > > met_help at ucar.edu> wrote:
> > >
> > > >
> > > > Thu Nov 21 08:09:37 2019: Request 93208 was acted upon.
> > > > Transaction: Ticket created by john.w.raby2.civ at mail.mil
> > > > Queue: met_help
> > > > Subject: METViewer v2.11 question
> > > > Owner: Nobody
> > > > Requestors: john.w.raby2.civ at mail.mil
> > > > Status: new
> > > > Ticket <Caution-Caution-Caution-Caution-url: Caution-
> > > Caution-Caution-
> Caution-https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=93208 >
> > > >
> > > >
> > > > When plotting continuous statistics ME and RMSE using MET
input .stat
> > > > files from Point-Stat, which settings in the "Statistics"
control
> panel
> > > of
> > > > the web GUI would generate statistics which match those
produced in
> the
> > > MET
> > > > text output file CNT.txt?
> > > >
> > > > Thanks.
> > > >
> > > > R/
> > > > John
> > > >
> > > > Mr John W. Raby, Meteorologist
> > > >
> > > > U.S. Army Research Laboratory
> > > >
> > > > White Sands Missile Range, NM 88002
> > > >
> > > > (575) 678-2004 DSN 258-2004
> > > >
> > > > FAX (575) 678-1230 DSN 258-1230
> > > >
> > > > Email: john.w.raby2.civ at mail.mil<Caution-Caution-Caution-
Caution-mailto:
> > john.raby at us.army.mil>
> > > >
> > > >
> > >
> > >
> > >
> >
> >
> >
> >
>
>
>
>
------------------------------------------------
More information about the Met_help
mailing list