R help - Oct 2012 - Problems plotting a sparse time series in R

Hi guys,

I am logging data about my body (weight, body fat, blood pressure, ..) 
in a .csv file and would like to plot this as a time series. I uploaded 
the (noisified) .csv, you can see the link in the code I have so far 
(you can run the code directly as-is):

     df.raw <- 
read.csv("http://www.chaotic-neutral.de/temp/stats-noised.csv",
sep=";",
dec=".")
     date <- as.Date(df.raw[,1])
     comments <- df.raw[,2]
     stats <- df.raw[,-c(1,2)]

     ################################################################
     #### timeSeries
     library(timeSeries)
     ts <- timeSeries(charvec=date, data=stats)
     plot(ts)

     ################################################################
     #### zoo
     library(zoo)
     ts <- zoo(stats, date)
     plot(ts)

I tried plotting the stuff with two packages, timeSeries and zoo. Both 
have the same problem: The column "Bauchfalte" which has a lot of 
entries (I measure this constantly), gets plotted, but the other ones 
don't. Most of them have only two entries and a lot of NAs, so I would 
at least expect a straight line between the two dates in those plots.

Maybe there is a problem with determining the ylim of the plots of the 
sparse columns, this is what an error message said. But I have no idea 
how to fix this.

Does anyone have an idea what's wrong here?

Also, if you have suggestions if there is a better format or way to log 
my data, I would appreciate that as well. I just figured the current way 
of one central date column and a lot of NAs for stuff I don't measure 
that day is the best way, because I'd like to add vertical lines at the 
dates with an entry in the "comment" column in every plot.

Thanks a lot!
  - Alex

On Mon, Oct 29, 2012 at 10:14 AM, Alexander Engelhardt
<alex at chaotic-neutral.de> wrote:
> Hi guys,
>
> I am logging data about my body (weight, body fat, blood pressure, ..) in a
> .csv file and would like to plot this as a time series. I uploaded the
> (noisified) .csv, you can see the link in the code I have so far (you can
> run the code directly as-is):
>
>     df.raw <-
> read.csv("http://www.chaotic-neutral.de/temp/stats-noised.csv",
sep=";",
> dec=".")
>     date <- as.Date(df.raw[,1])
>     comments <- df.raw[,2]
>     stats <- df.raw[,-c(1,2)]
>
>     ################################################################
>     #### timeSeries
>     library(timeSeries)
>     ts <- timeSeries(charvec=date, data=stats)
>     plot(ts)
>
>     ################################################################
>     #### zoo
>     library(zoo)
>     ts <- zoo(stats, date)
>     plot(ts)
>
Thanks for the reproducible example.
> I tried plotting the stuff with two packages, timeSeries and zoo. Both have
> the same problem: The column "Bauchfalte" which has a lot of
entries (I
> measure this constantly), gets plotted, but the other ones don't. Most
of
> them have only two entries and a lot of NAs, so I would at least expect a
> straight line between the two dates in those plots.
>
> Maybe there is a problem with determining the ylim of the plots of the
> sparse columns, this is what an error message said. But I have no idea how
> to fix this.
>
> Does anyone have an idea what's wrong here?
>
I'm not 100% certain, but I think plot.zoo only plots _consecutive_
observations as lines.
> Also, if you have suggestions if there is a better format or way to log my
> data, I would appreciate that as well. I just figured the current way of
one
> central date column and a lot of NAs for stuff I don't measure that day
is
> the best way, because I'd like to add vertical lines at the dates with
an
> entry in the "comment" column in every plot.
>
I would plot the sparse data as points, not lines.  For example:
plot(ts, screens=1, type=c("l",rep("p",4)), col=rainbow(5))
> Thanks a lot!
>  - Alex
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
Best,
--
Joshua Ulrich  |  about.me/joshuaulrich
FOSS Trading  |  www.fosstrading.com

On Oct 29, 2012, at 8:14 AM, Alexander Engelhardt wrote:
> Hi guys,
>
> I am logging data about my body (weight, body fat, blood  
> pressure, ..) in a .csv file and would like to plot this as a time  
> series. I uploaded the (noisified) .csv, you can see the link in the  
> code I have so far (you can run the code directly as-is):
>
>    df.raw <-
read.csv("http://www.chaotic-neutral.de/temp/stats-noised.csv
> ", sep=";", dec=".")
>    date <- as.Date(df.raw[,1])
>    comments <- df.raw[,2]
>    stats <- df.raw[,-c(1,2)]
>
>    ################################################################
>    #### timeSeries
>    library(timeSeries)
>    ts <- timeSeries(charvec=date, data=stats)
>    plot(ts)
>
>    ################################################################
>    #### zoo
>    library(zoo)
>    ts <- zoo(stats, date)
>    plot(ts)
If you instead use this:

ts <- zoo(stats, date)
plot(ts, type="p")

You may get a better idea why your other panels in plot.zoo are emply.  
Y\In them you have no adjacent points that are not missing. R plotting  
routines generally break aa line plot at a missing value rather than  
drawing a line through missing data.

-- 

David.
>
> I tried plotting the stuff with two packages, timeSeries and zoo.  
> Both have the same problem: The column "Bauchfalte" which has a
lot
> of entries (I measure this constantly), gets plotted, but the other  
> ones don't. Most of them have only two entries and a lot of NAs, so  
> I would at least expect a straight line between the two dates in  
> those plots.
>
> Maybe there is a problem with determining the ylim of the plots of  
> the sparse columns, this is what an error message said. But I have  
> no idea how to fix this.
>
> Does anyone have an idea what's wrong here?
>
> Also, if you have suggestions if there is a better format or way to  
> log my data, I would appreciate that as well. I just figured the  
> current way of one central date column and a lot of NAs for stuff I  
> don't measure that day is the best way, because I'd like to add  
> vertical lines at the dates with an entry in the "comment" column
in
> every plot.
>
> Thanks a lot!
> - Alex
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
David Winsemius, MD
Alameda, CA, USA