Providing custom labels (xticks) to line()

We are trying to experiment with bokeh coming from a Pandas & matplotlib environment. One of the things that’s missing is plotting series data with an x-axis which is an alphanumeric set of labels.

I don’t know what this is called in bokeh terminology, but matplotlib allows you set every xtick label explicitly like:

plt.xticks(range(len(xticks)), xticks)

Where xticks = [‘a’,‘b’,‘c’]…

The best analogous feature I could find was a FuncTickFormatter, but this requires you to provide JS (or Python) to perform the same function. Unfortunately, the example on the web shows you referring to a ‘tick’ parameter and there’s no apparent way to pass in data to this function, so I don’t have an easy way to accomplish this.

FYI, we are on bokeh 0.12.10 which is tied to the particular Anaconda release we have.

Thanks,

-Clint

Hi,

You question is a bit underspecified, it sounds like you have a continuous numeric axis (not a categorical axis), but that you want to label fixed positions in a custom way?

Then you can set .ticker on the axis to specify fixed tick positions:

  https://bokeh.pydata.org/en/0.12.10/docs/reference/models/axes.html#bokeh.models.axes.Axis.ticker

And you can set .major_label_overrides on the axis to map those fixed positions to arbitrary labels:
  
  https://bokeh.pydata.org/en/0.12.10/docs/reference/models/axes.html#bokeh.models.axes.Axis.major_label_overrides

Something like:

  p.xaxis.ticker = [ 1, 5, 10 ]
  p.xaxis.major_label_overrides = { 1: 'a', 5: 'b', 10: 'c' }

If you have actual categorical data and axes, you should refer to this chapter of the users guide:

  https://bokeh.pydata.org/en/0.12.10/docs/user_guide/categorical.html

Thanks,

Bryan

···

On Jul 25, 2018, at 15:04, Clint Olsen <[email protected]> wrote:

We are trying to experiment with bokeh coming from a Pandas & matplotlib environment. One of the things that's missing is plotting series data with an x-axis which is an alphanumeric set of labels.

I don't know what this is called in bokeh terminology, but matplotlib allows you set every xtick label explicitly like:

plt.xticks(range(len(xticks)), xticks)

Where xticks = ['a','b','c']...

The best analogous feature I could find was a FuncTickFormatter, but this requires you to provide JS (or Python) to perform the same function. Unfortunately, the example on the web shows you referring to a 'tick' parameter and there's no apparent way to pass in data to this function, so I don't have an easy way to accomplish this.

FYI, we are on bokeh 0.12.10 which is tied to the particular Anaconda release we have.

Thanks,

-Clint

--
You received this message because you are subscribed to the Google Groups "Bokeh Discussion - Public" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].
To post to this group, send email to [email protected].
To view this discussion on the web visit https://groups.google.com/a/continuum.io/d/msgid/bokeh/df24b1e5-d438-4d06-bcfd-b6c705acec12%40continuum.io.
For more options, visit https://groups.google.com/a/continuum.io/d/optout.

Hi:

I tried your change, and it works well. I used:

p.xaxis.major_label_overrides = dict(zip(range(len(xticks)), xticks))

p.xaxis.major_label_orientation = ‘vertical’

p.xaxis.ticker = list(range(len(xticks)))

Yes, this is categorical data. However, we like to use line plots so we can superimpose the results on the same chart. Bar charts don’t lend themselves as well to have many independent trials. You can think of the plots as connecting the tops of the bars to produce a behavioral curve.

Thanks,

-Clint

···

On Wednesday, July 25, 2018 at 3:15:51 PM UTC-7, Bryan Van de ven wrote:

Hi,

You question is a bit underspecified, it sounds like you have a continuous numeric axis (not a categorical axis), but that you want to label fixed positions in a custom way?

Then you can set .ticker on the axis to specify fixed tick positions:

    [https://bokeh.pydata.org/en/0.12.10/docs/reference/models/axes.html#bokeh.models.axes.Axis.ticker](https://bokeh.pydata.org/en/0.12.10/docs/reference/models/axes.html#bokeh.models.axes.Axis.ticker)

And you can set .major_label_overrides on the axis to map those fixed positions to arbitrary labels:

    [https://bokeh.pydata.org/en/0.12.10/docs/reference/models/axes.html#bokeh.models.axes.Axis.major_label_overrides](https://bokeh.pydata.org/en/0.12.10/docs/reference/models/axes.html#bokeh.models.axes.Axis.major_label_overrides)

Something like:

    p.xaxis.ticker = [ 1, 5, 10 ]

    p.xaxis.major_label_overrides = { 1: 'a', 5: 'b', 10: 'c' }

If you have actual categorical data and axes, you should refer to this chapter of the users guide:

    [https://bokeh.pydata.org/en/0.12.10/docs/user_guide/categorical.html](https://bokeh.pydata.org/en/0.12.10/docs/user_guide/categorical.html)

Thanks,

Bryan

On Jul 25, 2018, at 15:04, Clint Olsen [email protected] wrote:

We are trying to experiment with bokeh coming from a Pandas & matplotlib environment. One of the things that’s missing is plotting series data with an x-axis which is an alphanumeric set of labels.

I don’t know what this is called in bokeh terminology, but matplotlib allows you set every xtick label explicitly like:

plt.xticks(range(len(xticks)), xticks)

Where xticks = [‘a’,‘b’,‘c’]…

The best analogous feature I could find was a FuncTickFormatter, but this requires you to provide JS (or Python) to perform the same function. Unfortunately, the example on the web shows you referring to a ‘tick’ parameter and there’s no apparent way to pass in data to this function, so I don’t have an easy way to accomplish this.

FYI, we are on bokeh 0.12.10 which is tied to the particular Anaconda release we have.

Thanks,

-Clint


You received this message because you are subscribed to the Google Groups “Bokeh Discussion - Public” group.

To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].

To post to this group, send email to [email protected].

To view this discussion on the web visit https://groups.google.com/a/continuum.io/d/msgid/bokeh/df24b1e5-d438-4d06-bcfd-b6c705acec12%40continuum.io.

For more options, visit https://groups.google.com/a/continuum.io/d/optout.

Just to be clear, categorical factors in Bokeh are always strings, and a true categorical axis in Bokeh does not have the .ticker attribute all. I would describe your approach more or less as "faking" a categorial axis, not in pejorative way, but merely to be clear that the underlying axis and coordinates you are using are continuous numerical ones.

Thanks,

Bryan

···

On Jul 25, 2018, at 15:38, Clint Olsen <[email protected]> wrote:

Hi:

I tried your change, and it works well. I used:

    p.xaxis.major_label_overrides = dict(zip(range(len(xticks)), xticks))
    p.xaxis.major_label_orientation = 'vertical'
    p.xaxis.ticker = list(range(len(xticks)))

Yes, this is categorical data. However, we like to use line plots so we can superimpose the results on the same chart. Bar charts don't lend themselves as well to have many independent trials. You can think of the plots as connecting the tops of the bars to produce a behavioral curve.

Thanks,

-Clint

On Wednesday, July 25, 2018 at 3:15:51 PM UTC-7, Bryan Van de ven wrote:
Hi,

You question is a bit underspecified, it sounds like you have a continuous numeric axis (not a categorical axis), but that you want to label fixed positions in a custom way?

Then you can set .ticker on the axis to specify fixed tick positions:

        https://bokeh.pydata.org/en/0.12.10/docs/reference/models/axes.html#bokeh.models.axes.Axis.ticker

And you can set .major_label_overrides on the axis to map those fixed positions to arbitrary labels:
         
        https://bokeh.pydata.org/en/0.12.10/docs/reference/models/axes.html#bokeh.models.axes.Axis.major_label_overrides

Something like:

        p.xaxis.ticker = [ 1, 5, 10 ]
        p.xaxis.major_label_overrides = { 1: 'a', 5: 'b', 10: 'c' }

If you have actual categorical data and axes, you should refer to this chapter of the users guide:

        https://bokeh.pydata.org/en/0.12.10/docs/user_guide/categorical.html

Thanks,

Bryan
         
> On Jul 25, 2018, at 15:04, Clint Olsen <[email protected]> wrote:
>
> We are trying to experiment with bokeh coming from a Pandas & matplotlib environment. One of the things that's missing is plotting series data with an x-axis which is an alphanumeric set of labels.
>
> I don't know what this is called in bokeh terminology, but matplotlib allows you set every xtick label explicitly like:
>
> plt.xticks(range(len(xticks)), xticks)
>
> Where xticks = ['a','b','c']...
>
> The best analogous feature I could find was a FuncTickFormatter, but this requires you to provide JS (or Python) to perform the same function. Unfortunately, the example on the web shows you referring to a 'tick' parameter and there's no apparent way to pass in data to this function, so I don't have an easy way to accomplish this.
>
> FYI, we are on bokeh 0.12.10 which is tied to the particular Anaconda release we have.
>
> Thanks,
>
> -Clint
>
> --
> You received this message because you are subscribed to the Google Groups "Bokeh Discussion - Public" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].
> To post to this group, send email to [email protected].
> To view this discussion on the web visit https://groups.google.com/a/continuum.io/d/msgid/bokeh/df24b1e5-d438-4d06-bcfd-b6c705acec12%40continuum.io.
> For more options, visit https://groups.google.com/a/continuum.io/d/optout.

--
You received this message because you are subscribed to the Google Groups "Bokeh Discussion - Public" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].
To post to this group, send email to [email protected].
To view this discussion on the web visit https://groups.google.com/a/continuum.io/d/msgid/bokeh/5ac45e22-cce2-4d7d-9e7b-3793680efd66%40continuum.io.
For more options, visit https://groups.google.com/a/continuum.io/d/optout.