Skip to content

Exploratory Data Analysis (EDA)

Jump to bottom
Niko Aarnio edited this page Oct 28, 2024 · 6 revisions

Overview

The Exploratory Data Analysis (EDA) page facilitates data exploration with plots, statistics and exploratory tools in their own tabs.

image

Plots

Using EIS Wizard plots, you can visualize distributions and relationships between variables to identify (anomaly) thresholds, correlations and other patterns of interest. Currently the following plot types are available:

  • Histogram
  • KDE
  • ECDF
  • Scatterplot
  • Lineplot
  • Barplot
  • Boxplot
  • Pairplot (vector)
  • Pairplot (raster)
  • Parallel coordinates plot (vector)
  • Parallel coordinates plot (raster)

Some of the plot types are divided into separate pages for vector and raster data while some implement the plot type for both data types in the same page. Note that currently some plot types are not implemented for raster data.

To produce a plot/graph, first select the desired plot type from the available plots listed above (1). When the plot type is changed, the data selection and parameter widgets are automatically updated to match the selected plot type.

After the plot type is set, select your data (2). The supported data types are automatically filtered, so the drop-down menu will only show layers that can be used for the plot type. For some plots, you need define only one feature or raster to plot, for other multiple are required.

The Plot parameters section (3) consists of some generic parameters, such as color and opacity, and several plot-specific parameters. You can use the parameters to customize the plot to your liking, but using them is optional.

A plot can be generated by pressing the Create plot button. The plot appears below in a space reserved for it. Should you want to inspect the plot in a larger window, you can press the Open in new window button which creates a popup window with the plot. Note that this will transfer the plot to the new window, so it will disappear/get broken in the embedded view. For further adjustments on the plot and saving it to a file (in new window), you can use the navigation bar above the plot.

The Reset button resets only the plot parameters to their defaults and does not affect the produced plot, plot type or input data. The Clear plot button will remove the produced plot from the embedded plot view. Note that clicking Clear plot will also close the popup window with your plot if you have transferred it to a new window.

EDA_histogram

Statistics

You can compute various statistics on both vector and raster data. This helps you to get an overview of the data you wish to use in your Mineral prospectivity mapping (MPM) workflow and find thresholds to define anomalies. The statistics consist of general and descriptive statistics, and quantiles:

  • General statistics
    • N total
    • N NULL
  • Descriptive statistics
    • Mean
    • Standard deviation
    • Relative standard deviation
    • Variance
    • Skewness
  • Quantiles
    • Min
    • 25 % quantile
    • Median
    • 75 % quantile
    • Max

Start by selecting a layer from the dropdown menu (1). Select a field or band (2) from the selected layer for which you want to compute the statistics. You can adjust the number of decimals to be reported (up to 10) using the Decimals spinbox (2). Compute the statistics by pressing the Compute button (3). The results will then appear in the empty slots (4).

EDA_statistics

Exploratory analysis

This view lists EIS tools available under EDA category. No custom user interfaces are implemented (yet), but links are displayed to the processing algorithms.

Clone this wiki locally