Will the sun shine? - An accessible dataset for teaching machine learning and deep learning

Florian Huber, Erica Dafne van Kuppevelt, Peter Steinbach, Colin Sauze, Yang Liu, Berend Weel

Abstract: Hands-on teaching of modern machine learning and deep learning techniques heavily relies on the use of well-suited datasets. We present a novel tabular dataset that was specifically created for teaching machine learning and deep learning to an academic audience. The dataset contains intuitively accessible weather observations from 18 locations in Europe. It was designed to be suitable for a large variety of different training goals, and to avoid reaching unrealistically high prediction accuracy. Teachers or instructors thus can choose the difficulty of the training goals and thereby match it with the respective learner audience or lesson objective.