Many of our final products that we deliver to our clients are what might initially appear to be ‘simple maps’. These spatial models combine and condense a large amount of complex information about the distribution of species, ecological processes, ecosystem status and threatening processes to represent key ideas in simple spatial forms. Although the products may first appear simple, there is an enormous amount of information embedded within them.
The data that we use can come from many different families or types of information, and these may not initially or naturally combine together. As an example, we may want to know more about the likely distribution of a rare plant. The rarity of the species may infer that quality information or observations of this species may be very limited. Integrating this data directly with other content-rich data sources such as satellite imagery, digital terrain models or climate data may not the best approach to making a useful product. Considerable thought is required when combining data together, so that the most appropriate and informative methods are used to deliver the best possible product. These insights and skills are a major strength of the research group.
Much of our work involves using tools called machine learning algorithms. These complex mathematical approaches allow the computer to solve the problem. This sounds simple and easy ….. I can hear you say “ great!, just let the computer do the hard work ” ! But it is not necessarily that simple. Considerable thought needs to be given to the structure and coding of the information so that the computer works on the problem or questions that you are interested, and not on some other ‘problem’.