Statistical Analysis

  • Statistical Analysis
  • Data Selection
  • Univariate Statistics
  • Univariate Statistics Plots
  • Bivariate Statistics
  • Trivariate Statistics
  • Multivariate Statistics

Statistical Analysis

Maptek has recently completed its Statistics II module which is now available. This module allows input from multiple data sources in any particular run. This means that data from ASCII files, block models, drill hole and compositing databases can be analysed in one single analysis and data from various sources compared to each other. The various graphs requested are posted next to each other in user defined styles and colours, ready for plotting next to raws or block slices. The Geostatistics toolkit also provides a comprehensive set of statistical tools.

Data Selection

Data selection is possibly the most critical part of any particular evaluation. Vulcan supports many modes of data selection, apart from the actual data source selection. Once the particular variables for analysis have been chosen, the selection of applicable values must be specified. This selection includes the application of conditions (0.0 < au > 100), the selection of specific values by numeric tags, ignoring specific values (-99) and including only values in certain ranges.

Selection can also be made by applying character tags. In addition to these selections, it is also possible to demarcate data by choosing data lying inside or outside of a particular (or set of) triangulation. This overcomes the problem of selecting data from different rock types in the case of a straight composite.

Univariate Statistics

Univariate statistics can be computed and displayed in various forms, depending on the task at hand. By default the statistics module calculates a standard set of statistics for each variable selected. These statistics are:

  • The number of samples used
  • Minimum and maximum values
  • Range
  • Average
  • Median
  • Standard deviation
  • Variance
  • Geometric mean and variance
  • Harmonic mean
  • Skewness
  • Kurtosis
  • Mean of the natural logs
  • Variance of the natural logs
  • Sichels T estimate
  • D1 and D3 statistics

Univariate Statistics Plots

Vulcan supports the following univariate plots:

  • General statistics
  • Histogram/line plots with user definable intervals and data display
  • Pie charts - user definable
  • Cumulative distribution
  • Lognormal probability plots
  • Character histograms

The plots of histograms and pie charts can be plotted with either scale defined as linear, logarithmic or normal Gaussian. Third parameter log-normal distributions can also be calculated. Confidence limits and the grades above various cut-off values can also be easily calculated.

Bivariate Statistics

In addition to the statistics mentioned above, Vulcan supports five types of bivariate statistical plots and analysis. These are:

  • Joint plots - plots of correlation coefficients between two variables
  • Scatter - scatter plots of two variables with the correlation coefficients calculated
  • Line - plots of sequential data pairs
  • QQ plots - distribution comparisons by value
  • PP plots - distribution comparisons by percentiles

Trivariate Statistics

Trivariate statistics are displayed using ternary plots. Vulcan grid calculator also supports 3D ternary plots.

Multivariate Statistics

Here it is possible to calculate various plots as defined above for numerous variables and perform them in a single run. For example, it is possible to calculate a histogram, general statistics and a scatter plot for each of copper, nickel and iron from a composite file in one single run.