Hello, I am trying to create a plot (I guess two plots) using two different time series datasets, but I'm not sure of the best approach. The data is from a tidal surge due to a hurricane and I would like to show the relationship between stage and windspeed/direction. One dataset is the tidal stage and it is sampled in 30 minutes intervals. The other dataset is windspeed and direction and it is sampled every 6 minutes. I would like to display the tidal hydrograph and then also show windspeed and direction as a banner above it. They would be two separate plots, but with identical x-axis (as opposed to one plot that has both tide & wind.) Should I combine the two time series into one dataframe (with lots of NAs) so that I can create the plots together, for example using pairs or ggpairs? Or should I keep them separate, make the individual plots, and then work on the layout in graphic design software? I would prefer the former, but I just can't figure out how to make that work out. Also, is it possible to depict wind direction (which is given in degree clockwise from due north) with an arrow? Thank you very much, Ezra -- Ezra Boyd, PhD DisasterMap.net, LLC <http://DisasterMap.net> ezgis <ezgis76 at gmail.com>76 at gmail.com <ezgis76 at gmail.com> (504)533-4447 [[alternative HTML version deleted]]
Jeff Newmiller
2015-Nov-06 01:25 UTC
[R] Attempting to plot two different time series together
Seems like you would benefit from reading about long and wide data... perhaps
[1].
If I am understanding what you want, ggplot facetting should be able to do what
you want. You first have to put the data in long form (e.g. variable,
timestamp, value) before you give it to ggplot.
If you want more specific help then you should provide a small sample data set
as one or more R dputs as the Posting Guide recommends (see footer, and [2]).
[1] https://cran.r-project.org/web/packages/tidyr/vignettes/tidy-data.html
[2]
http://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example
---------------------------------------------------------------------------
Jeff Newmiller The ..... ..... Go Live...
DCN:<jdnewmil at dcn.davis.ca.us> Basics: ##.#. ##.#. Live
Go...
Live: OO#.. Dead: OO#.. Playing
Research Engineer (Solar/Batteries O.O#. #.O#. with
/Software/Embedded Controllers) .OO#. .OO#. rocks...1k
---------------------------------------------------------------------------
Sent from my phone. Please excuse my brevity.
On November 5, 2015 7:00:18 AM PST, Ezra Boyd <ezgis76 at gmail.com>
wrote:>Hello,
>
>I am trying to create a plot (I guess two plots) using two different
>time
>series datasets, but I'm not sure of the best approach. The data is
>from a
>tidal surge due to a hurricane and I would like to show the
>relationship
>between stage and windspeed/direction.
>
>One dataset is the tidal stage and it is sampled in 30 minutes
>intervals.
>The other dataset is windspeed and direction and it is sampled every 6
>minutes. I would like to display the tidal hydrograph and then also
>show
>windspeed and direction as a banner above it. They would be two
>separate
>plots, but with identical x-axis (as opposed to one plot that has both
>tide
>& wind.)
>
>Should I combine the two time series into one dataframe (with lots of
>NAs)
>so that I can create the plots together, for example using pairs or
>ggpairs? Or should I keep them separate, make the individual plots, and
>then work on the layout in graphic design software? I would prefer the
>former, but I just can't figure out how to make that work out.
>
>Also, is it possible to depict wind direction (which is given in degree
>clockwise from due north) with an arrow?
>
>Thank you very much,
>
>Ezra
Jeff Newmiller
2015-Nov-06 16:36 UTC
[R] Attempting to plot two different time series together
Please keep the list in the conversation.
Yes stacking was my intention, since the graphical presentation does not need
the same time basis. However, your other analyses may indeed require that you
interpolate or decimate to obtain aligned data records.
---------------------------------------------------------------------------
Jeff Newmiller The ..... ..... Go Live...
DCN:<jdnewmil at dcn.davis.ca.us> Basics: ##.#. ##.#. Live
Go...
Live: OO#.. Dead: OO#.. Playing
Research Engineer (Solar/Batteries O.O#. #.O#. with
/Software/Embedded Controllers) .OO#. .OO#. rocks...1k
---------------------------------------------------------------------------
Sent from my phone. Please excuse my brevity.
On November 6, 2015 8:01:55 AM PST, Ezra Boyd <ezgis76 at gmail.com>
wrote:>Jeff,
>
>Thank you very much for responding. I actually read and worked through
>most of ggplot2 and ggplot2 Essentials for this task, but was starting
>to
>think they were not what I needed.
>
>The problem that I cannot get over is that my two timeseries datasets
>are
>based on different samplings, 30-mins for the tidal and 6-mins for the
>meteorological:
>
>Tidal Data
>TS Stage_ft 1/9/2001 16:00 -0.41 1/9/2001 16:30 -0.43 1/9/2001 17:00
>-0.44 1/9/2001
>17:30 -0.43 1/9/2001 18:00 -0.4
>Meteorological
>DATETIME WINDSPEED DIR 8/15/2012 0:00 5.05 228 8/15/2012
>0:06
>4.08 225 8/15/2012 0:12 4.08 216 8/15/2012 0:18 4.47 222 8/15/2012 0:24
>5.05
>219 8/15/2012 0:30 3.5 226
>
>Do I understand your suggestion correctly that I should tidy up the
>dataset
>by stacking them in the dataframe, and then facet them using a type
>column?
>TS Stage_ft WINDSPEED DIR Type 1/9/2001 16:00 -0.41 T 1/9/2001
>16:30
>-0.43 T 1/9/2001 17:00 -0.44 T 1/9/2001 17:30 -0.43 T 1/9/2001
>18:00 -0.4 T 8/15/2012 0:00 5.05 228 M 8/15/2012 0:06 4.08 225
> M 8/15/2012
>0:12 4.08 216 M 8/15/2012 0:18 4.47 222 M 8/15/2012 0:24 5.05
>219 M 8/15/2012
>0:30 3.5 226 M
>Or, do I need to look into timeseries methods for combining datasets
>with
>different samplings?
>
>Thanks again and I appreciate you taking the time to share your
>knowledge.
>
>Regards,
>
>Ezra
>
>
>
>
>
>
>
>
>On Thu, Nov 5, 2015 at 7:25 PM, Jeff Newmiller
><jdnewmil at dcn.davis.ca.us>
>wrote:
>
>> Seems like you would benefit from reading about long and wide data...
>> perhaps [1].
>>
>> If I am understanding what you want, ggplot facetting should be able
>to
>> do what you want. You first have to put the data in long form (e.g.
>> variable, timestamp, value) before you give it to ggplot.
>>
>> If you want more specific help then you should provide a small sample
>data
>> set as one or more R dputs as the Posting Guide recommends (see
>footer,
>> and [2]).
>>
>> [1]
>https://cran.r-project.org/web/packages/tidyr/vignettes/tidy-data.html
>>
>> [2]
>>
>http://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example
>>
>---------------------------------------------------------------------------
>> Jeff Newmiller The ..... ..... Go
>Live...
>> DCN:<jdnewmil at dcn.davis.ca.us> Basics: ##.#.
##.#. Live
>> Go...
>> Live: OO#.. Dead: OO#..
>Playing
>> Research Engineer (Solar/Batteries O.O#. #.O#. with
>> /Software/Embedded Controllers) .OO#. .OO#.
>rocks...1k
>>
>---------------------------------------------------------------------------
>> Sent from my phone. Please excuse my brevity.
>>
>> On November 5, 2015 7:00:18 AM PST, Ezra Boyd <ezgis76 at
gmail.com>
>wrote:
>> >Hello,
>> >
>> >I am trying to create a plot (I guess two plots) using two
different
>> >time
>> >series datasets, but I'm not sure of the best approach. The
data is
>> >from a
>> >tidal surge due to a hurricane and I would like to show the
>> >relationship
>> >between stage and windspeed/direction.
>> >
>> >One dataset is the tidal stage and it is sampled in 30 minutes
>> >intervals.
>> >The other dataset is windspeed and direction and it is sampled
every
>6
>> >minutes. I would like to display the tidal hydrograph and then
also
>> >show
>> >windspeed and direction as a banner above it. They would be two
>> >separate
>> >plots, but with identical x-axis (as opposed to one plot that has
>both
>> >tide
>> >& wind.)
>> >
>> >Should I combine the two time series into one dataframe (with lots
>of
>> >NAs)
>> >so that I can create the plots together, for example using pairs or
>> >ggpairs? Or should I keep them separate, make the individual plots,
>and
>> >then work on the layout in graphic design software? I would prefer
>the
>> >former, but I just can't figure out how to make that work out.
>> >
>> >Also, is it possible to depict wind direction (which is given in
>degree
>> >clockwise from due north) with an arrow?
>> >
>> >Thank you very much,
>> >
>> >Ezra
>>
>>
Jeff -- Thank you for the clarification. Ezra On Fri, Nov 6, 2015 at 10:36 AM, Jeff Newmiller <jdnewmil at dcn.davis.ca.us> wrote:> Please keep the list in the conversation. > > Yes stacking was my intention, since the graphical presentation does not > need the same time basis. However, your other analyses may indeed require > that you interpolate or decimate to obtain aligned data records. > --------------------------------------------------------------------------- > Jeff Newmiller The ..... ..... Go Live... > DCN:<jdnewmil at dcn.davis.ca.us> Basics: ##.#. ##.#. Live > Go... > Live: OO#.. Dead: OO#.. Playing > Research Engineer (Solar/Batteries O.O#. #.O#. with > /Software/Embedded Controllers) .OO#. .OO#. rocks...1k > --------------------------------------------------------------------------- > Sent from my phone. Please excuse my brevity. > > On November 6, 2015 8:01:55 AM PST, Ezra Boyd <ezgis76 at gmail.com> wrote: > >Jeff, > > > >Thank you very much for responding. I actually read and worked through > >most of ggplot2 and ggplot2 Essentials for this task, but was starting > >to > >think they were not what I needed. > > > >The problem that I cannot get over is that my two timeseries datasets > >are > >based on different samplings, 30-mins for the tidal and 6-mins for the > >meteorological: > > > >Tidal Data > >TS Stage_ft 1/9/2001 16:00 -0.41 1/9/2001 16:30 -0.43 1/9/2001 17:00 > >-0.44 1/9/2001 > >17:30 -0.43 1/9/2001 18:00 -0.4 > >Meteorological > >DATETIME WINDSPEED DIR 8/15/2012 0:00 5.05 228 8/15/2012 > >0:06 > >4.08 225 8/15/2012 0:12 4.08 216 8/15/2012 0:18 4.47 222 8/15/2012 0:24 > >5.05 > >219 8/15/2012 0:30 3.5 226 > > > >Do I understand your suggestion correctly that I should tidy up the > >dataset > >by stacking them in the dataframe, and then facet them using a type > >column? > >TS Stage_ft WINDSPEED DIR Type 1/9/2001 16:00 -0.41 T 1/9/2001 > >16:30 > >-0.43 T 1/9/2001 17:00 -0.44 T 1/9/2001 17:30 -0.43 T 1/9/2001 > >18:00 -0.4 T 8/15/2012 0:00 5.05 228 M 8/15/2012 0:06 4.08 225 > > M 8/15/2012 > >0:12 4.08 216 M 8/15/2012 0:18 4.47 222 M 8/15/2012 0:24 5.05 > >219 M 8/15/2012 > >0:30 3.5 226 M > >Or, do I need to look into timeseries methods for combining datasets > >with > >different samplings? > > > >Thanks again and I appreciate you taking the time to share your > >knowledge. > > > >Regards, > > > >Ezra > > > > > > > > > > > > > > > > > >On Thu, Nov 5, 2015 at 7:25 PM, Jeff Newmiller > ><jdnewmil at dcn.davis.ca.us> > >wrote: > > > >> Seems like you would benefit from reading about long and wide data... > >> perhaps [1]. > >> > >> If I am understanding what you want, ggplot facetting should be able > >to > >> do what you want. You first have to put the data in long form (e.g. > >> variable, timestamp, value) before you give it to ggplot. > >> > >> If you want more specific help then you should provide a small sample > >data > >> set as one or more R dputs as the Posting Guide recommends (see > >footer, > >> and [2]). > >> > >> [1] > >https://cran.r-project.org/web/packages/tidyr/vignettes/tidy-data.html > >> > >> [2] > >> > > > http://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example > >> > > >--------------------------------------------------------------------------- > >> Jeff Newmiller The ..... ..... Go > >Live... > >> DCN:<jdnewmil at dcn.davis.ca.us> Basics: ##.#. ##.#. Live > >> Go... > >> Live: OO#.. Dead: OO#.. > >Playing > >> Research Engineer (Solar/Batteries O.O#. #.O#. with > >> /Software/Embedded Controllers) .OO#. .OO#. > >rocks...1k > >> > > >--------------------------------------------------------------------------- > >> Sent from my phone. Please excuse my brevity. > >> > >> On November 5, 2015 7:00:18 AM PST, Ezra Boyd <ezgis76 at gmail.com> > >wrote: > >> >Hello, > >> > > >> >I am trying to create a plot (I guess two plots) using two different > >> >time > >> >series datasets, but I'm not sure of the best approach. The data is > >> >from a > >> >tidal surge due to a hurricane and I would like to show the > >> >relationship > >> >between stage and windspeed/direction. > >> > > >> >One dataset is the tidal stage and it is sampled in 30 minutes > >> >intervals. > >> >The other dataset is windspeed and direction and it is sampled every > >6 > >> >minutes. I would like to display the tidal hydrograph and then also > >> >show > >> >windspeed and direction as a banner above it. They would be two > >> >separate > >> >plots, but with identical x-axis (as opposed to one plot that has > >both > >> >tide > >> >& wind.) > >> > > >> >Should I combine the two time series into one dataframe (with lots > >of > >> >NAs) > >> >so that I can create the plots together, for example using pairs or > >> >ggpairs? Or should I keep them separate, make the individual plots, > >and > >> >then work on the layout in graphic design software? I would prefer > >the > >> >former, but I just can't figure out how to make that work out. > >> > > >> >Also, is it possible to depict wind direction (which is given in > >degree > >> >clockwise from due north) with an arrow? > >> > > >> >Thank you very much, > >> > > >> >Ezra > >> > >> > >-- Ezra Boyd, PhD DisasterMap.net, LLC <http://DisasterMap.net> ezgis <ezgis76 at gmail.com>76 at gmail.com <ezgis76 at gmail.com> (504)533-4447 [[alternative HTML version deleted]]