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.

[ Procedures ] [ Concepts ] [ Exploratory Analysis ] [ Semivariogram Analysis ]
[ Semivariogram Models ] [ Model Validation ] [ Kriging ]


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