Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Create Vega specification for AUROC bar plots #84

Merged
merged 4 commits into from
Feb 20, 2017

Conversation

patrick-miller
Copy link
Member

This introduces a Vega based AUROC bar plot, used for comparing models on different mutations.

image

I'm thinking we maybe add a dashed line for the test partition performance of the covariates-only model on each of these to give a sense for a baseline?

@dhimmel
Copy link
Member

dhimmel commented Feb 14, 2017

@patrick-miller really nice! Looks like the test/train color scheme is similar to the ROC curve plot, which is great -- let's make the colors exactly the same if possible!

I'm thinking we maybe add a dashed line for the test partition performance of the covariates-only model on each of these to give a sense for a baseline?

Yeah, I think a dashed line could work. IIn some cases the covariate AUROC could be greater than the full AUROC. Accordingly, I think this is actually a difficult visualization that may take some trial and error to see what works. So I'd try out what you think will communicate this effectively!

@patrick-miller
Copy link
Member Author

Here is my first attempt at adding covariate only results:

image

The color scheme was the same, I had just added some transparency. I included the same transparency here.

@dhimmel
Copy link
Member

dhimmel commented Feb 16, 2017

Nice. I think we should show the covariate performance for both testing and training. Is it possible to split the red dash into two? You could even use the same color scheme (orange / purple) for the dashses without the transparency.

@patrick-miller
Copy link
Member Author

Cool, how does this look -- is it clear enough?

image

Copy link
Member

@dhimmel dhimmel left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@patrick-miller that's great. I like how you can see the covariate performance, but it's not visually overwhelming.

@dhimmel dhimmel merged commit 63cd3f9 into cognoma:master Feb 20, 2017
@patrick-miller patrick-miller deleted the feature-auroc_barplot branch February 20, 2017 16:40
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

2 participants