Skip to content

Plot MArginal distribution in the correct place

8 messages · Pedro páramo, Jim Lemon

Pedro páramo
# 
Hi all,

I have this code, but the marginal distribution plot doesn?t appear aligned
with the left plot.


I think could be something about layout or par() mar.

The code was programmed by me time ago.

Can anyone help me to get the marginal distribution on the center (more
higher centered)

id.txt

Could have this code:

05/01/2016;9335,200195
06/01/2016;9197,400391
07/01/2016;9059,299805
08/01/2016;8909,200195
11/01/2016;8886,099609
12/01/2016;8915,400391
13/01/2016;8934,5
14/01/2016;8787,700195
15/01/2016;8543,599609
18/01/2016;8469,299805
19/01/2016;8554,900391
20/01/2016;8281,400391
21/01/2016;8444,200195
22/01/2016;8722,900391
25/01/2016;8567,700195
26/01/2016;8692,5
27/01/2016;8741



g<-read.table("id.txt", col.names=c("Dateh","LAST"), sep=";", dec=",")

N=5000
B=24
ghy<-nrow(g)
r<-as.numeric(as.character(g$LAST[ghy]))


nf<-layout(matrix(c(1,1,1,1,2,2),1,6,byrow=TRUE))

par(mar=c(6,6,6,0.5))

A<-matrix(1:B,B,N);



sigma<-0.06;



mu<-0.00;


Z<-r*exp((mu-0.5*((sigma)^2)*A) +sigma*(sqrt(A))*matrix( rnorm(N*B,0,1), B,
N))

real1<-g$LAST[1:nrow(g)]

real2<-matrix(NA,nrow(g),N-1)

real<-cbind(real1,real2)




Po<-r*matrix(1,1,N);



Sim<-rbind(Po,Z)
Simulation<-rbind(real,Z)






par(mar=c(10,6,6,6))
matplot(Simulation,type="l",ylim=c(0,40000))

abline(h = 8000, lwd = 2, col = "black")

abline(h = 12000, lwd = 2, col = "black")
title("Dinamic Montecarlo Simulation 2 years ahead",font=4)

fhist<-hist(Simulation,plot=FALSE)
par(mar=c(6,0,6,6))
barplot(fhist$counts,axes=FALSE, space=0,horiz=TRUE,col="lightgray")
grid()
title("Marginal Distribution",font=4)


rect(0, 0, 0, 0) # transparent
Pedro páramo
# 
Please, I'm trying to put the right plot higher and centered on the left
values but I don't achive.

I would appreciate so much your help

El 6 jun. 2017 22:37, "Pedro p?ramo" <percentil101 at gmail.com> escribi?:

  
  


  


      

    

    

      
      

        


  

        

      Jim Lemon
    

    

      
      #
    

  


  

      
Hi Pedro,
As a one-off, you just shove the coordinates around a bit:

par(mar=c(11,0,6,6))
barplot(fhist$counts,axes=FALSE, space=0,horiz=TRUE,col="lightgray",
 ylim=c(0,24))

However, I don't think that this plot illustrates quite what you think it does.

Jim

        
On Wed, Jun 7, 2017 at 4:01 PM, Pedro p?ramo <percentil101 at gmail.com> wrote:

            

      
      
      
    

        

      
      
    

  
  


  


      

    

    
1 day later

    

      
      

        


  

        

      Pedro páramo
    

    

      
      #
    

  


  

      
Many thanks Jim.

What I,m trying to show with the fhist plot is the empirical distribution
of the values of the left plot simulation.

You say:
However, I don't think that this plot illustrates quite what you think it
does.

Can you give me a clue to try to illustrate better if it is not showing
what I believe it shows a better way to show it?

Many thanks in advance.





El 7 jun. 2017 12:08, "Jim Lemon" <drjimlemon at gmail.com> escribi?:

Hi Pedro,
As a one-off, you just shove the coordinates around a bit:

par(mar=c(11,0,6,6))
barplot(fhist$counts,axes=FALSE, space=0,horiz=TRUE,col="lightgray",
 ylim=c(0,24))

However, I don't think that this plot illustrates quite what you think it
does.

Jim

        
On Wed, Jun 7, 2017 at 4:01 PM, Pedro p?ramo <percentil101 at gmail.com> wrote:

            

      
      
      
    

        

      
      
    

        
posting-guide.html

        

      
      
    

        

  
  


  


      

    

    
5 days later

    

      
      

        


  

        

      Pedro páramo
    

    

      
      #
    

  


  

      
Please can you send me some orientation?

Many thanks in advance.

Only if posible one book o similar example to understand why it is not what
I try.

El 8 jun. 2017 7:50 PM, "Pedro p?ramo" <percentil101 at gmail.com> escribi?:

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

  
  


  


      

    

    

      
      

        


  

        

      Jim Lemon
    

    

      
      #
    

  


  

      
Hi Pedro,
If you keep that same margins for the second plot:

par(mar=c(10,0,6,6))
barplot(fhist$counts,axes=FALSE, space=0,horiz=TRUE,col="lightgray")

it looks reasonably well aligned to me. Because you are plotting the
counts of the values in Simulation, the ordinate (vertical axis) of
the bar plot is in quite different units from that of the plot on the
left side.

Jim

        
On Wed, Jun 14, 2017 at 5:33 PM, Pedro p?ramo <percentil101 at gmail.com> wrote:

            

      
      
      
    

        

      
      
    

            

      
      
      
    

  
  


  


      

    

    
3 days later

    

      
      

        


  

        

      Pedro páramo
    

    

      
      #
    

  


  

      
Can you suggest a plot on the right side so that the right plot has the
same units of left plot and reflect the counts of the number of time the
values repeat, something like the empirical distribution?



2017-06-14 12:30 GMT+02:00 Jim Lemon <drjimlemon at gmail.com>:

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

  
  


  


      

    

    

      
      

        


  

        

      Jim Lemon
    

    

      
      #
    

  


  

      
As you have categorized the original values, they are different from
the originals. You could include the categories like this, with a bit
of fiddling the margins and y limits for the right hand plot:

x11(width=10)
nf<-layout(matrix(c(1,1,1,1,2,2),1,6,byrow=TRUE))
par(mar=c(10,6,6,6))
matplot(Simulation,type="l",ylim=c(0,40000))
abline(h = 8000, lwd = 2, col = "black")
abline(h = 12000, lwd = 2, col = "black")
title("Dinamic Montecarlo Simulation 2 years ahead",font=4)
fhist<-hist(Simulation,plot=FALSE)
par(mar=c(10,6,6,6))
barplot(fhist$counts,axes=FALSE, space=0,horiz=TRUE,col="lightgray",
 names.arg=paste(fhist$breaks[1:23],fhist$breaks[2:24],sep="-"),
 las=2,ylim=c(-3,40))
grid()
title("Marginal Distribution",font=4)

Jim

        
On Sun, Jun 18, 2017 at 3:51 AM, Pedro p?ramo <percentil101 at gmail.com> wrote: