Plotting on the right axis in Bokeh

Question

I'm building a multi-categorial boxplot and I'm having trouble aligning the boxplots to be in the right position.

I have my x axis of genes which is subdivided by databases. But my plot plots everything in the middle on gene.

I make my axis like this:

x = sorted(list(set([(gene, database) for gene in totaldf['gene'].to_list() for database in totaldf['db'].to_list()])))
p = figure(background_fill_color="#efefef", x_range=FactorRange(*x), width=1600, height=700)

and I try to plot my boxes like this:

p.vbar(x=(gene, db), width=0.7, bottom=q2[gene, db], top=q3[gene, db], fill_color="#ffffff", line_color="black")
p.vbar(x=(gene, db), width=0.7, bottom=q1[gene, db], top=q2[gene, db], fill_color="#ffffff", line_color="black")

This results in my plot plotting like this: https://imgur.com/a/8Q8YC1N

How do I get the plot to be in the right locations? The dataframe looks like this:

      gene       db  mutations
0     IGHV1-3  G1K_CL2          6
1    IGHV1-58  G1K_CL2          2
2    IGHV1-58  G1K_CL2          3
3     IGHV1-8  G1K_CL2          2
4    IGHV3-16  G1K_CL2          3
..        ...      ...        ...
141  IGHV4-61  G1K_CL3         11
142  IGHV4-61  G1K_CL3         12
143  IGHV4-61  G1K_CL3         10
144  IGHV4-61  G1K_CL3         13
145  IGHV7-81  G1K_CL3          4

Answer 1

gene and db are the columns from the DataFrame? The coordinates are not being supplied in the correct format. You are supplying a 2-tuple of lists , but what is required is a list of 2-tuples . The coordinate list should look like:

x=[(gene1, db1), (gene2, db2), ...])

Probably zip(gene, db) will provide what you intend.

All that said, I also strongly advise to use explicitly created ColumnDataSource when dealing with nested categorical data. There are some inherent ambiguities that can arise and constructing a CDS yourself eliminates those.