Data-Driven Modeling Features

Software-OverviewThe Data-Driven Modeling Module is the essential extension for any geological modeling. It provides a novel way in the industry to create geological and block models based on drawing workflows. The users sets up a workflow of tasks to perform on the data and presses the run button to calculate the results. The is a very flexible yet powerful solution that will allow you to be up and running quickly while still being capable of handling complex tasks.

GeoReka!‘s Data-Driven Modeling Module provides an extensive range of tools to create a variety of models. Depending on the available data start with extracting information from your drilling, followed by converting the extracted data to boundary surfaces and subsequently one or multiple models. Or for a quick first pass interpolation connect your assay data directly to one of the numeric interpolation algorithms.

Feature extraction

As a first step for most models information needs to be extracted from drilling data. Generally this means extracting points. This can be done either semi-automatically by defining what unit to extract or using a more manual approach. The manual approach allows you to draw selection boxes around the drilling data in the Viewer to include or exclude parts of the data for further processing. GeoReka! can extract boundary points between layers or create soft boundaries using numeric data like alterations, grades or geochemistry.

Geological modeling

layered-filled-model-sideNearly all geological features are represented as surfaces at some stage in the modelling process in GeoReka!. It is therefore not a surprise that the Data-Driven Modeling Module contains several ways to create surfaces from points efficiently. In addition, there are a number of tools to aid in representing the geological features in a optimal way. These include smoothing, reducing the number of triangles or even re-triangulating existing surfaces.

Fault models

Faults can be created in various ways. The simplest is a flat plane created directly from the Section Plane in the Viewer. Slightly more advanced is the option to extrude a polyline to a surface with a customizable dip and azimuth. More advanced options use implicit modelling techniques to extrapolate points and polylines to a fault representation with an arbitrary resolution. Faults can be used to create domains which in turn can constrain numeric and geological models.


Domains in GeoReka! are solids created from a model extent and one or more surfaces. As mentioned, a domain can be used to constrain subsequent geological and numeric models, but can also be used to extract subsets of the data. These options are helpful for large scale projects that need to be broken down into smaller regions or mineralization that is constrained.

3D-veinOptimally oriented geological models

One of the unique abilities of geological modelling in GeoReka! is that models can be arbitrarily oriented. If your model needs to be dipping along a certain strike it can be done. This can greatly reduce your modelling times and memory usage. It is especially useful for thin laminar structures like veins, dykes, coal seams and other thin beds.

Thin Structures: veins, dykes, laminar beds

As mentioned above, GeoReka! can model thin laminar structures. The most versatile approach will allow for manual selection of intervals to be included into a model. This applies to numeric, e.g. grades, and category data, e.g. lithology. Using a distance from points as a boundary can even add some reliability indication into it. More automated ways allow for extraction of all hanging wall or footwall points and use those to create boundary surfaces.

Numeric interpolation, i.e. block modelling

3D-MA-ShellGeoReka!‘s Data-Driven Modeling Module has multiple interpolation techniques ranging from the trusted Inverse Distance Weighting to more advanced Radial Basis Functions (RBFs), commonly referred to as Implicit Modeling and even experimental techniques like Neural Networks. For dense data sets the type of interpolation will not make a huge difference. For more sparse data sets determining the areas of uncertainty will be a high priority. Using a distance from points indication can help, but being able to compare results from different techniques using the same data and range settings will greatly contribute to defining drill targets. GeoReka! currently provides the following methods: Inverse Distance Weighting, Moving Average, RBF and Neural Network. For additional control and direction all except the Inverse Distance Weighting can use a Global Anisotropy setting.

Save and re-use modeling workflows

Workflows created in the Data-Driven Modeling Module can be saved to be re-used at a later stage. This allows you to develop your own workflow that fits your project. That way if new data is available or changes to a data set have been made, the workflow can be loaded again and connected to the new data. A simple press of a button will start crunching out your new model(s).

Audit trails

One of the many advantages of the Data-Driven Modeling Module is the ability to generate a quick overview of the steps involved to obtain a certain outcome. Use screenshots of the workflows to explain and highlight how models were created. As well, the ability to save a workflow means that the parameter settings can be easily retained.