This public work investigates how modern machine-learning methods can be used to produce Digital Soil Mapping (DSM) products that support statistical inference, both at local scale (individual locations) and at higher aggregation levels such as land-use classes. Some of the key drivers were the following questions: Which statistical quantities must be estimated for DSM maps…
Integrated Process Modeling Software
An Integrated Process Model (IPM) within the pharmaceutical industry leverages on the entirety of process knowledge (i.e. both experimental and manufacturing data) from all the different steps to quantitatively understand and predict how changes affect the final product. As a relatively novel technique it is used as part of process validation for setting up a…
Real estate analyser
Immolyser tool is developed for real estate analysis. Among others, it helps in answering the following questions: Home value estimate. Our model uses insights from the Zillow’s Home Value Prediction competition. Drivers for real estate prices. What-If scenario simulation. The optimal price under constraint of (most) certain sale within X weeks. Similar estates based on user…
Anomaly detection in time-series
For a client we developed tailor-made algorithms for anomaly detection in air and hydrostatic pressure time-series. These algorithms were documented and implemented in gwloggeR, a R package made publicly available on github. To make the results understandable, the package also output diagnostic plots such as these: