In this thesis we will examine architectures and models for machine learning in three problem domains each of which are based around the use of time series data in time series applications. We set out to examine whether the architecture and model solutions in different problem domains will converge when optimised towards a similar solution or not. Stated clearly, our central research question is “That problem-solving in diverse problem domains using Machine Learning applied to time series data requires diverse models in order to achieve the best performance” . To investigate this research hypothesis we use a case study methodology. We will investigate three separate and diverse problem domains, and compare their results and best solutions. The first problem domain is in the field of educational analytics, the second is in the field of agri-analytics and the third is in the field of environmental science.