We have written a series of articles on working with geographic data, based on a project we conducted. The project aimed to predict areas where wolves will settle. This article provides an overview of the entire project, including links to the individual steps we took during the implementation of this project. You can find them at the bottom of this page. We also have a scientific article published about this project.

After being gone for more than 150 years, the wolf is once again part of the Dutch fauna today. After the first wolf was spotted in the Netherlands in 2015, the wolf resettled permanently in the Netherlands three years later. A year later - in 2019 - it was time for beschuit met muisjes, when the first cubs were born in the Netherlands. So we can no longer ignore it: the wolf is back!

With the return of the wolf, the debate over whether the wolf belongs here has flared up considerably. Proponents argue that the wolf is indigenous, and that its return is important for ecosystem dynamics. Opponents worry because wolves kill livestock and because the wolf is increasingly being spotted in inhabited areas.

Whether you are happy about the wolf or not, the fact remains that the wolf is back. A great opportunity for us to get to work with geographic data for a change. To see if we can predict where the wolf will settle, we went back to the origin of the wolf in the Netherlands: Germany. Can we predict for wolves in Germany where they will settle?

To answer this question, we looked at where wolves had already settled in Germany. By looking at different environmental characteristics, we tried to identify likely areas where new wolves would settle. The environmental characteristics we looked at included the type of landscape, the presence of infrastructure, population density and the presence of other wolves and other species.

The most important predictor of wolf settlement is the presence of another wolf, in this regard the distance to these other wolves does play an important role. Wolves are territorial, so a new pair will only establish a territory outside the territory of already present wolves. But again, the distance to other wolves should not be too great, because without other wolves there is no possibility to reproduce.

What is interesting is that our analysis has shown that population density seems to play a lesser role than previously thought. Of course, this does not immediately mean that we expect wolves to settle in the center of a city, but we do see that the amount of forestation and the presence of other wolves plays a larger role in determining settlement than population density.

Below is a map of Germany showing where wolves are currently established and where wolves are likely to become established in the future.

Thus, based on this data, we see a clear pattern in which places wolves settle. This can help predict where wolves will settle in Germany in the future. So for now we have only looked at Germany, but since wolves do not adhere to national boundaries you could assume that these factors are also important in other countries for the likelihood that a wolf pair will be present.

If you are interested in how we approached this analysis, be sure to read our scientific article, or the other articles in this series:

  1. Getting value out of geodate with AI: getting started
  2. Getting value out of geodate with AI: convert locations to their lat and lon
  3. Getting value out of geodate with AI: data preparation
  4. Getting value out of geodate with AI: train the model
  5. Getting value out of geodate with AI: explainability using SHAP
  6. Getting value out of geodate with AI: visualize the model predictions

If you want to get started on this project, you can also go directly to the Python code toe.

 

Recent posts

Automating sales proposals with AI

Automating sales proposals with AI

Discover how AI improves sales proposal automation. Using Microsoft Copilot Studio, we streamline workflows and increase efficiency by integrating AI-driven tools. Learn more about the benefits and challenges of this innovative approach.

read more