I am working on a basic Neural network and want to show performance of model with respect to different parameters.Here is my dataframe looks like.

I need help with suggestions for impactful visualizations. I can not make countourplots as far as I understand because for that I have to define loss as a function of the parameters. Please suggest some interesting plots.

  • $\begingroup$ Simple visualisations can be effective, so I suggest plotting the parameter of interest against the performance using a line plot or bar chart. $\endgroup$
    – Lynn
    Commented Sep 25, 2022 at 12:20

1 Answer 1


You can use a "Parallel Coordinates" plot, in which

  • A model is represented by a line
  • Model performance is highlighted in colour
  • Model hyper-parameters are presented in axes along with their respective values.

I attach below an example which you can generate either using Tensorboard or tools like Weights & Biases.

You can find here and here resources on how to implement this representation for each of the tools above, respectively.

Hope this helps, good luck!

enter image description here


Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.