Species Distribution Modelling (SDM, also known as Habitat Distribution Modelling) is a framework that allows reconstructing the potential range of a species based on its occurrences and environmental factors of interest. It is often used to inform habitat restoration (for example, by taking into account future climatic scenarios) and ecological planning (e.g. for invasive species).

Tidysdm reaches a whole new level of flexibility compared with existing tools. This is achieved using the modular infrastructure of tidymodels. Tidymodels is a collection of R packages for machine learning, in which syntax, grammar and data structure are fully compatible with each other.

This is why tidysdm does not need to create complete solutions from scratch: objects created within tidysdm can be directly fed to existing generic functions from other tidymodels packages. In addition to it, it provides metrics specific to SDM and functions to handle spatial data.

Tidysdm is the first available piece of software to perform SDMs on time series. This is a task that is often achieved by splitting the observations into different time slices, but such an approach reduces the data used for each modelling step and may lead to biases.

Finally, tidysdm allows full integration with pastclim, an R package facilitating the access to present-day climatic data and future scenarios (e.g. Worldclim, CHELSA), as well as palaeoclimatic reconstructions covering from thousands to millions of years.

In short, tidysdm allows full flexibility when performing SDM, making it suitable to inform a larger range of ecological questions, and can natively incorporate time-scattered data, facilitating the study of climatic changes (especially the ones we are currently experiencing) on both large and small geographic scales.