Back to the Topics Page

Predicting local urban temperatures using crowd-sourced data and machine-learning

Objective:

In cities, temperatures are often higher than in the surrounding areas and can also vary locally. This can depend on the type and color of surfaces, as well as the morphology. This "city effect" is called the urban heat island (UHI). We want to understand the climate variability at a local scale, as it can have a large impact on natural systems. We don't have a full theoretical understanding of these complex interactions yet. Therefore we propose to use a combination of a dense(r), crowdsourced, low-cost sensor network and machine learning techniques, to collect more data and find interactions between a huge number of variables.

Keywords:

Description:

Problem definition

The goal is to predict temperatures locally at street level. To determine the city temperatures this research focuses on crowdsourced data from the WOW-NL and OpenSenseMap network. In contrast to the 33 official ground-based observations from the KNMI, the WOW-NL and OpenSenseMap networks measure at more than 300 locations within cities and outside cities. Initially we want to predict the temperatures at the same location as the measurements. Additional information from other data sources about the stations' location, like position, population density, elevation, sky view factor, and exposure will be used as input for the machine learning algorithms.

Methodology

The research starts with a literature study to gain more insight in city temperatures at street level and low-cost sensor networks. Which variables are of importance, and how reliable are the measurements? Then, the focus will be on machine learning algorithms using R or Python libraries. We propose to use libraries that use allow to explore several machine learning techniques such as: neural network, regression tree, support vector machine, and random forest. For these techniques pre-processing of the data is often required, using for example scaling, centering, box-cox transformations and principle component analysis. In the second phase of the research the focus is on ensemble predictions: using multiple machine learning techniques to predict street temperatures.

Expected results

From the student we expect a scientific report including literature background and research results. Besides a report we also aim at documentation and reproducibility of code using repositories such as Github.

Important organizational note

This research topic is offered in collaboration with the Royal Netherlands Meteorological Institute (KNMI). Formal supervision will be carried out by an ITC staff member, but daily supervision on content will be together with Dr. Irene Garcia Marti from KNMI. Supervision details will be discussed and agreed on prior to starting the thesis research.

References:

Domain(s):

Study Program(s):

Researchers working on this field:

Printer-friendly version
Send by email