Spatial Predictive Modelling
Spatial predictive modelling is a powerful tool that can be widely applied. It provides a way of integrating digital spatial data and specialised knowledge using GIS modelling techniques to effectively target business activities and operations.
Possible applications of predictive modelling include:
- Agricultural - e.g. locating the most productive land for farming
- Horticultural - e.g. modelling of ideal climate and soil conditions for growth
- Conservation - e.g. finding rare occurrences of endangered species
- Environmental - e.g. identifying land parcels that may be of significant environmental importance
- Geological - e.g. locating the most prospective areas for mineral exploration
- Geotechnical risk - e.g. identifying areas prone to landslides or flooding
- Emergency services - e.g. areas of particular hazard risk during earthquakes or storms
With large amounts of data now freely available in a digital format along with vast amounts of data collected in-house we recognise a problem for organisations of data overload. We can help by compiling, managing, and analysing the data in a comprehensive GIS and spatial predictive modelling system.
Beginning with a customised GIS database, Kenex adds value to the data using spatial data modelling techniques developed by the Canadian Geological Survey. The techniques employed include: weights of evidence, fuzzy logic, and logistic regression. These techniques take different layers of information that have been derived from the original data and combine them using statistical methods to produce an integrated result that predicts the probability of a variable in question occurring.
Ranked targets generated from the model results can be used to create a portfolio with the most technically, commercially or environmentally viable sites, reducing the development costs and timeframe of a project.
The development of a predictive model by Kenex includes the following steps:
Data compilation: All data relevant to the model needs to be collected in a digital form. In many cases data is already available in digital GIS format either free-of-charge or through a licence agreement. Kenex organises the digitising of the remaining data. We combine data that we collect for a specific project with data provided by the client into a single GIS database that will be used for the predictive modelling. This stage includes a complete QC of the data to make sure it is the best quality available.
Expert knowledge: No model could be completed without expert knowledge of the process being modelled. This knowledge can come from in-house if it is with one of our areas of expertise, from the client, or from a contracted expert.
Spatial Analysis: Datasets are tested using spatial statistics and are converted into the key predictive maps required for the model. This stage is the most time consuming and requires scores of different filtering, analytical, statistical, data preparation and analysis techniques.
Modelling:Kenex uses Arc-SDM on an ESRI GIS platform and MI-SDM on a MapInfo GIS platform to run the models and prepare the predictive maps. As one of the global leaders in predictive modelling we maintain close links with the software developers to ensure that we're all constantly developing the modelling techniques and the science behind them.
Model Output and Targeting: The modelling process generates an output layer that shows the probability of the event occurring at a given location. This dataset can be reclassified to display the most prospective areas, it could be applied to a set of land parcels showing which are most likely to have the event modelling occurring in it, or used to rank the importance of existing spatial objects e.g. mineral permits or land use zones. A list of ranked targets can be generated that are attributed with the model parameters.
- Final Reporting and Delivery: The reporting and delivery is tailored to the client's specifications. These can range from a complete digital dataset in the form of a GIS through to a set of maps showing the model results. Kenex provides a report for all our models discussing the results, how they were obtained, and what data was included.