search

exhuma / python-cluster

Simple clustering library for python.
GNU Lesser General Public License v2.1
65 stars 27 forks source link

Proper visualisation of clusters #24

Closed 1kastner closed 7 years ago

1kastner commented 7 years ago

I currently visualize the output of hierarchical clustering this way:

cluster = HierarchicalClustering(
    [custom_time_series_object_1, custom_time_series_object_2, ...],
    lambda x, y: compare_time_series(x, y, "temperature")
)
cluster.cluster()
cluster._data[0].display()

Somehow this feels a bit dirty as the underscore indicates that it is a private member. Is there a better, more sophisticated way how to do this? I somehow want to have an informative dendogram.

exhuma commented 7 years ago

I am currently unaware of a cleaner way to accomplish this. I agree that it should be a public method on the top-level object. Something like cluser.display() in your example.

I'll flag it as enhancement.

For now, don't worry using is as such. I currently have not a lot of spare time to work on this project so the functionality is unlikely to disappear. I think a good solution would be to simply implement the "display" function as a simple delegate to the function you are using right now. I'll get to it once I have more time. Unless you want to submit a PR ;)

1kastner commented 7 years ago

Closed with #25

exhuma commented 7 years ago

By the way, I also pushed it to pypi as 1.4.0.

Thanks for the PR :)

1kastner commented 7 years ago

Oh nice :)

On 05/27/2017 09:56 AM, Michel Albert wrote:

By the way, I also pushed it to pypi as 1.4.0.

Thanks for the PR :)

— You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub https://github.com/exhuma/python-cluster/issues/24#issuecomment-304436223, or mute the thread https://github.com/notifications/unsubscribe-auth/AE_lxZzzWbIten_q8tYV1bLcChrznEwzks5r99dJgaJpZM4NiEh8.