I have heatmap data (the x and y axis take in angles as inputs, and the color of that particular spot on the grid represents the "divergence.") I am unsure on how to numerically describe my data (not including descriptive statistics). I want to be able to quantify clusters/the presence of clusters on my heatmaps. I experimented with cluster heatmaps (with dendrograms on the axis), but I am unsure on what to conclude from these heatmaps (how many clusters, where they are located, how distinct these clusters are, etc.) I want to be able to extrapolate some numerical quantifiers for my data. I am happy to provide more information. Thanks in advance.

EDIT: I am looking at sensitivity to initial conditions for a double pendulum. I measure the divergence in trajectories (with some small perturbation of the angles) after a given amount of time; divergence is the euclidean distance (angle_1, angle_2, velocity_1, velocity_2) of the phase space. I am happy to provide my code if needed.

(Here are the heatmaps I was referring to, sorry if the scaling is bad): heatmap of my data cluster heatmap of my data

  • $\begingroup$ What physical experimental setup produced the (angle1, angle2, divergence) readings? How is divergence measured or defined? It would really help if you would Cite your Reference(s) in the text of the question. $\endgroup$
    – J_H
    Commented Feb 24 at 6:20


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.