A brief set of analyses for a subset of users of a banking app that aims to save people money automatically using AI.
It begins with some simple exploratory analysis, such as histograms, scatter plots, and bar charts, to identify visually if there are any interesting correlations and to have an intuitive sense of the data and therefore the users.
Then K-Means, an ML algorithm, is implemented to cluster the users in a self-supervised manner. This revealed a noticeable divide between high-saving and low-saving users. It would be interesting to push this further and develop more meaningful clusters of users. This could be achieved through increasing k past 4, feeding in different data from the users, or combining/computing data points (i.e. feature engineering).