Challenge
AM is a software company based in Germany hiring me on a contract basis when they want something of quality and delivered fast. This new project was for a client of theirs who wanted to build a recommendation system for tailored investements plans. One of the slight complexity was to build a decision tree algorithm which could be easily updated by the client. The other one was not resorting to bring ML libraries as AM wanted React/Javascript.
Solution
I am a machine learning enthusiast and further improving on the field so by already understanding how decision tree works, this facilitated development. What is a decision tree ? A decision tree is a supervised machine learning algorithm that can be used for both Regression and Classification problem statements. Hence, by remembering that the final output of the Decision Tree is a Tree having Decision nodes and leaf nodes, I managed to reproduce the logic with Javascript without having to create any Machine learning models with Python SciPy.
The design was ready on Figma which facilited my job to just transforming the design to pixel perfect code. The outcome is working with the precision of a Swiss watch and a UI which match that last statement.