The purpose of the thesis was to compare open-source systems that provide tools to build neural networks. Selected systems for the thesis were Knime and Ml5.js library. To help in comparison, neural networks were built with both systems. Theory of this thesis was based on literature and websites about the thesis subject. Also, documentation from the systems used in the thesis were used. In the theory part of the thesis goal was to give basic understanding of how neural networks and what are their possibilities. In functional part covers neural network build process with both systems. The thesis was functional. Hockey player stats were used to teach neural networks. Goal with the neural networks was to predict players next seasons points. Results indicated that there is lot of similarities and the be found in how they can be used. Result obtained from the neural networks had only slight difference in them. Ml5.js works great as a light and easy to use add-on for websites. Knime on the other hand works great in more analysis-based work, thanks to its good extension possibilities.