Stocks price trend point-cloud for custom time intervals
This post is the continuation of a previous post made about machine learning and stocks (here) . To understand the concepts that will be discussed, it is necessary to understand those discussed earlier. I want to continue with the idea of representing stock price trends (STPs) using points in a cartesian graph. Last time, the SPT I had discussed were studied over a fixed period. However, it would be interesting to study how those SPTs evolve for different time periods.
To illustrate the use of Stock Price Trend (SPT) point-cloud, I have selected 8 stocks, 4 pharmaceutical ones and 4 related to oil companies. Depending on the time-interval selected I can guess which trends are closer to which other.

Multiple things can be inferred from the graph, first, pharmaceutical SPTs tend to be closer to each other than Oil SPTs; Their trends tend to be less diverse that those of Oil SPTs. Second, for a lot of time intervals, both types of SPTs seem to be at different sections of the canvas, indicating a clear distinction between the two types SPTs.