Time series plots or cluster numbers on wrong places

[ALindblomA]

Really nice with a package that gives the time series besides the dendrogram!
However, I encountered a problem both with the new update of the code and the old, i find that the timeseries plotted besides the dendrogram are not placed on the right ”spots”. I looked at the number for the clusters when i used allclust_draw and compared with the time series and they some how got mixed up, dont know if it is the timeseries plots which are placed wrong or the "names/positions". Will attach pictures of how it looks and how it should be according to my time series and their positions in the df.

Best regards
Anna :)

[mapattacker]

this is quite a critical error, let me find time this weekend to debug.

Appreciate the notice :)

[mapattacker]

hi @ALindblomA , can u pip install dendrogram-ts==0.1.2 and try again?

It should work for allclust_draw() function which you are using now.

[ALindblomA]

Hi, it seams to works now! Both for allclust and when usig the maxclust with equal amounts of clusters that I have timeseries. Thank you for fixing it so quickly.

Is there a way to instead of having the "positions" of my time series, put labels as their names from the column header? I have tried this when just using a normal dendrogram and it seams to work. but i guess it would have to bee put into the code some how?

Best regards
Anna

[mapattacker]

Hi Anna,

that's weird, I only managed to fix the allclust_draw function actually.

The other two I might need to ask the scipy community whether there's a bug in their code as I can't get it to display correctly. Hence likely require more time to resolve.

For the labels, can u post that as a new issue? Do u meant replace the cluster names with ur own label names? An example will help alot (like what is the normal dendrogram output u did that seem to work?)

[ALindblomA]

Oh okay, yes you are correct I was in a bit of a hurry yeasterday so I managed to run the allclust two times instead of the maxclust also, sorry.
I will create a new issue for the labels :)
/ Anna