Operational Spatial Machine Learning

Building operational machine learning models is a challenge - it covers multiple steps which involve very different technical skill, with the actual training of ML only being a small part. When you want to work with spatial data, your problems get yet another dimension.

We know that all of this is annoying. That is why we have developed a machine learning model cycle built on top of the UDL data warehouse that will immensely simplify all of that. With contextual spatial data waiting in an analysis-ready form, all you need to do is to link them to your data using a streamlined API. Our machine learning cycle takes care of the training, and a data warehouse covering the whole country at a high resolution allows prediction from one side of the nation to the other. The model is then available as an API as well as an interactive online map. When the new batch of data comes in, the whole pipeline is ready to generate an updated, more precise version of the model.

Your data analysts and data scientists, now how a sleek way to use spatial data for prediction and instantly make use of their results - including the sharing with to non-technical users in their team.

All that sounds great, right? Soon, you will be able to play with that! The UDL MLStudio will be available as a no-code online interface and a Python one for those who prefer to code.

Want a sneak peek right now? DM us ;-).
  #machinelearning #data #python #urbananalysis #urbanmorphology