R help - Nov 2011 - Bar charts, frequencies known, intervals of varying width

Hi

I would like to plot bar charts in a particular way. The intervals are not
evenly distributed and are in a data frame column called size and the
relative frequencies are in a second column called mass. Both size and mass
are continuous ratio data to all intents and purposes. The data actually
represents sieving of sand where $size is sieve aperture and where $mass is
the amount remaining on the sieve. 

So if I sieve 100g of sand through increasingly fine sieves nothing remains
on the coarsest sieve say 2500 microns. Between 2000 and 2500 microns there
is 0.2g remining on the 2000micron sieve etc. as per the data supplied.

size, mass
2500, 0
2000, 0.2
1700, 1.8
1400, 5
1000, 13
800, 40
600, 20
300, 15
180, 3
<180, 2

The data here I have made up for simplicity. In general sands tend to be
distributed as lognormal curves and mixtures of sands are combined as the
mixture of 2 or more lognormal distributions.

I would like to fit 2, 3 or preferably 4 parameter log normal distributions
and plot these cumulative distributions on the charts. 

The problems are: I can't seem to find how to do the bar charts in R and 
I have not found in the past how to fit 3 or 4 parameter log normal
cumulative distributions in R

Please help.
Thanks
Scriptham

--
View this message in context:
http://r.789695.n4.nabble.com/Bar-charts-frequencies-known-intervals-of-varying-width-tp4100114p4100114.html
Sent from the R help mailing list archive at Nabble.com.

Stewart,

I am replying to the graphical part of your query.

## This was constructed with
##       dump("scriptham2","")
## sso it can be copied directly from the email into an R session.


scriptham2 <-
structure(list(size = c(2500, 2000, 1700, 1400, 1000, 800, 600,
300, 180, 100), mass = c(0, 0.2, 1.8, 5, 13, 40, 20, 15, 3, 2
)), .Names = c("size", "mass"), row.names = c(NA, 10L),
class "data.frame")


xyplot(mass ~ size, data=scriptham2,
       scales=list(x=list(log=TRUE)),
       horizontal=FALSE, panel=panel.barchart, box.width=.05, origin=0)


xyplot(mass ~ size, data=scriptham2,
       scales=list(x=list(log=TRUE)),
       type="b", pch=16)

I think you are asking for the first graph.  I also think that a line
graph, as in the second graph,
is more appropriate for this type of data.

I arbitrarily assigned the "<180" to 100, as it has to have a
numerical
value on the x-axis.

Rich
On Wed, Nov 23, 2011 at 10:50 AM, scriptham2
<stewart.porteous@permarock.com
> wrote:
> Hi
>
> I would like to plot bar charts in a particular way. The intervals are not
> evenly distributed and are in a data frame column called size and the
> relative frequencies are in a second column called mass. Both size and mass
> are continuous ratio data to all intents and purposes. The data actually
> represents sieving of sand where $size is sieve aperture and where $mass is
> the amount remaining on the sieve.
>
> So if I sieve 100g of sand through increasingly fine sieves nothing remains
> on the coarsest sieve say 2500 microns. Between 2000 and 2500 microns there
> is 0.2g remining on the 2000micron sieve etc. as per the data supplied.
>
> size, mass
> 2500, 0
> 2000, 0.2
> 1700, 1.8
> 1400, 5
> 1000, 13
> 800, 40
> 600, 20
> 300, 15
> 180, 3
> <180, 2
>
> The data here I have made up for simplicity. In general sands tend to be
> distributed as lognormal curves and mixtures of sands are combined as the
> mixture of 2 or more lognormal distributions.
>
> I would like to fit 2, 3 or preferably 4 parameter log normal distributions
> and plot these cumulative distributions on the charts.
>
> The problems are: I can't seem to find how to do the bar charts in R
and
> I have not found in the past how to fit 3 or 4 parameter log normal
> cumulative distributions in R
>
> Please help.
> Thanks
> Scriptham
>
> --
> View this message in context:
>
http://r.789695.n4.nabble.com/Bar-charts-frequencies-known-intervals-of-varying-width-tp4100114p4100114.html
> Sent from the R help mailing list archive at Nabble.com.
>
> ______________________________________________
> R-help@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<http://www.r-project.org/posting-guide.html>
> and provide commented, minimal, self-contained, reproducible code.
>
	[[alternative HTML version deleted]]

On 11/24/2011 02:50 AM, scriptham2 wrote:
> Hi
>
> I would like to plot bar charts in a particular way. The intervals are not
> evenly distributed and are in a data frame column called size and the
> relative frequencies are in a second column called mass. Both size and mass
> are continuous ratio data to all intents and purposes. The data actually
> represents sieving of sand where $size is sieve aperture and where $mass is
> the amount remaining on the sieve.
>
> So if I sieve 100g of sand through increasingly fine sieves nothing remains
> on the coarsest sieve say 2500 microns. Between 2000 and 2500 microns there
> is 0.2g remining on the 2000micron sieve etc. as per the data supplied.
>
> size, mass
> 2500, 0
> 2000, 0.2
> 1700, 1.8
> 1400, 5
> 1000, 13
> 800, 40
> 600, 20
> 300, 15
> 180, 3
> <180, 2
>
> The data here I have made up for simplicity. In general sands tend to be
> distributed as lognormal curves and mixtures of sands are combined as the
> mixture of 2 or more lognormal distributions.
>
> I would like to fit 2, 3 or preferably 4 parameter log normal distributions
> and plot these cumulative distributions on the charts.
>
> The problems are: I can't seem to find how to do the bar charts in R...
sands<-read.csv("sands.csv")
library(plotrix)
barp(sands$mass,names.arg=sands$size,staxx=TRUE)

Jim