Exploratory Analysis - Geostatistics
This module shows how to perform an exploratory data analysis using univariate
and bivariate statistics. The univariate statistics shows how to organize and
synthesize a set of values, which is performed, mainly, through histograms.
Important histogram characteristics are organized in three groups:
-
Localization measures: average, minimum value, inferior quartile,
median, superior quartile and maximum value;
-
Dispersion measures:
variance and standard deviation;
-
Shape measures: asymmetry coefficient, kurtosis coefficient and
variation coefficient.
The bivariate statistics shows how to describe the relationship between two
variables, that is, the relationship between two data sets or two
distributions. This relationship can be visualized through a dispersion diagram (Scatter
Plot). The linear relationship degree between two variables can be
measured using the correlation coefficient.
Executing Exploratory
Analysis:
-
activate using the "Control Panel" the numerical model
IL which has samples of the spot heights type;
-
click on Analysis - Geostatistics - Exploratory Analysis ...;
-
in the "Exploratory Analysis" window click on the down arrow in the
Statistics box and select an item from the list ( Descriptive
Statistics, Histogram, Normal
Probability Plot or Scatter Plot) ;
-
click on Apply... . A new window will be presented,
depending on the option selected. See the details below.
Descriptive Statistics
Analysis:
-
in the "Exploratory Analysis" window select the Descriptive
Statistics
item;
-
click on Apply... . The "Data Report"
window is presented with the following data:
-
Number of Samples
-
Number of Valid Samples
-
Average
-
Variance
-
Standard Deviation
-
Variation Coefficient
-
Asymmetry Coefficient
-
Kurtosis Coefficient
-
Minimum Value
-
Inferior Quartile
-
Median
-
Superior Quartile
-
Maximum Value
-
it is possible to save the data in a ASCII text file.
Histogram Analysis:
-
in the "Exploratory Analysis"
window select the Histogram item;
-
click on Apply... .
The "Histogram" window is presented;
-
by default the histogram is presented using 10 buckets (or classes), using the
yellow color, and the Gaussian distribution, with average and standard
deviation computed from the sampling data, using the red color;
-
if you want to change the number of buckets, click on Number of Classes and
select a value from 1 to 20, and next click on
.
In a similar way, to visualize the normal probability plot, see
the next explanation.
Analyzing the Normal
Probability Plot:
-
in the "Exploratory Analysis" window select the Normal
Probability Plot
item;
-
click on Apply... . The "Normal Probability Plot"
window is presented;
-
by "default" the plot presented using the blue color is related to
sample data and the one using the red color is related to the Gaussian
distribution;
-
if you want to change the window size just, click and drag the window
boundaries.
Analyzing the Scatter Plot:
-
in the "Exploratory Analysis" window select the Scatter Plot item;
-
click on Select other IL... . The " Categories
and Infolayers" window is presented, where it is possible to select
a second numerical model IL;
-
click on Apply... .
The "Scatter Plot" window is presented;
-
if you want to change the window size just, click and drag the window
boundaries.
See also:
SPRING Spatial Analysis
Importing Data
Regionalized Variables
How to Apply? - Semivariogram Analysis
|