Scatter Plots

This script is meant to demonstrate the various possibilities offered by gstlearn and gstlearn.plot for calculating and representing scatter plots or h-scatter plots.

We first define a data base containing (nech) isolated points randomly located. The samples belong to a square of mesh equal to 1.

nech = 100
db = Db_createFillRandom(nech, 2, 0)

Representing the contents of the data base

ggplot() + plot(db) + plot.decoration(title="Data Set")

We simulate two random variables linked by a joint model

model = Model(nvar=2,ndim=2)
err = model$addCovFromParam(ECov_EXPONENTIAL(),range=0.8,sills=c(2,1,1,2))
err = model$addCovFromParam(ECov_EXPONENTIAL(),range=0.2,sills=c(1.1,-1,-1,1.1))
ggplot() + plot.model(model, ivar=0, jvar=1, hmax=1)

err = simtub(NULL,db, model)
db

## 
## Data Base Characteristics
## =========================
## 
## Data Base Summary
## -----------------
## File is organized as a set of isolated points
## Space dimension              = 2
## Number of Columns            = 5
## Total number of samples      = 100
## 
## Variables
## ---------
## Column = 0 - Name = rank - Locator = NA
## Column = 1 - Name = x-1 - Locator = x1
## Column = 2 - Name = x-2 - Locator = x2
## Column = 3 - Name = Simu.1 - Locator = z1
## Column = 4 - Name = Simu.2 - Locator = z2

Scatter Plot

In this section, we present the scatter plot, represented in two different manners. On each figure, we represent the first bissector (in red) and the regression line (in blue).

• as a set of isolated points

res = correlationPairs(db, db, "Simu.1", "Simu.2")

ggplot() + plot.correlation(db, "Simu.1", "Simu.2", asPoint=TRUE, 
                    flagBiss=TRUE, flagSameAxes=TRUE, flagRegr=TRUE)

• as cells (of a fictitious grid) painted with color representing point density

ggplot() + plot.correlation(db, "Simu.1", "Simu.2", asPoint=FALSE, 
                    flagBiss=TRUE, flagSameAxes=TRUE, flagRegr=TRUE, bins=20)

H-Scatter plot

In this section, we represent samples distant by a given distance. This distance is defined using the VarioParam description and selecting the lag of interest.

We first define the VarioParam set of calculation parameters: essentially, we define the lag and the number of lags.

varioparam = VarioParam_createOmniDirection(npas=10, dpas=0.1)

We represent the H-Scatter plot:

• as a set of isolated symbols

ggplot() + plot.hscatter(db, "Simu.1", "Simu.2", varioparam, ipas=8, asPoint=TRUE, 
                 flagBiss=TRUE, flagSameAxes=TRUE)

• as cells (of a fictitious grid) painted with color representing point density

ggplot() + plot.hscatter(db, "Simu.1", "Simu.2", varioparam, ipas=1, asPoint=FALSE, 
                    flagBiss=TRUE, flagSameAxes=TRUE, bins=20)