Statistical glyphs like in ggplot2/ggvis?

Are there any plans for statistical glyphs for exploratory data analysis? I’m thinking ggplot stuff like smooth loess, kernel density plots, regressions etc

Hi datnamer,

This is absolutely the plan. We have been working on clarifying the "charts" interface. This has involved some measure of experimentation to see what "fits" so we have not gone crazy adding new chart types while there is still this code churn. However, I we are now quite close to having a very stable charts API to build upon, and then we can start churning out all kinds of new chart types with confidence.

Thanks,

Bryan

···

On Apr 15, 2015, at 5:36 PM, datnamer <[email protected]> wrote:

Are there any plans for statistical glyphs for exploratory data analysis? I'm thinking ggplot stuff like smooth loess, kernel density plots, regressions etc

--
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/e92687c7-59c2-4863-9a65-5855e7067288%40continuum.io.
For more options, visit https://groups.google.com/a/continuum.io/d/optout.

Sounds awesome! Would these chart types be compos-able and layer-able so we can mix and match as well?

···

On Thursday, April 16, 2015 at 10:36:37 AM UTC-4, Bryan Van de ven wrote:

Hi datnamer,

This is absolutely the plan. We have been working on clarifying the “charts” interface. This has involved some measure of experimentation to see what “fits” so we have not gone crazy adding new chart types while there is still this code churn. However, I we are now quite close to having a very stable charts API to build upon, and then we can start churning out all kinds of new chart types with confidence.

Thanks,

Bryan

On Apr 15, 2015, at 5:36 PM, datnamer [email protected] wrote:

Are there any plans for statistical glyphs for exploratory data analysis? I’m thinking ggplot stuff like smooth loess, kernel density plots, regressions etc


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/e92687c7-59c2-4863-9a65-5855e7067288%40continuum.io.

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

Yes there is actually an open PR that implements some of that, that will be in the 0.9 release next month. (In fact that's exactly the sort of thing we want to "get right" before cranking on lots of new chart types)

Thanks,

Bryan

···

On Apr 16, 2015, at 10:18 AM, datnamer <[email protected]> wrote:

Sounds awesome! Would these chart types be compos-able and layer-able so we can mix and match as well?

On Thursday, April 16, 2015 at 10:36:37 AM UTC-4, Bryan Van de ven wrote:
Hi datnamer,

This is absolutely the plan. We have been working on clarifying the "charts" interface. This has involved some measure of experimentation to see what "fits" so we have not gone crazy adding new chart types while there is still this code churn. However, I we are now quite close to having a very stable charts API to build upon, and then we can start churning out all kinds of new chart types with confidence.

Thanks,

Bryan

> On Apr 15, 2015, at 5:36 PM, datnamer <[email protected]> wrote:
>
> Are there any plans for statistical glyphs for exploratory data analysis? I'm thinking ggplot stuff like smooth loess, kernel density plots, regressions etc
>
> --
> 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/e92687c7-59c2-4863-9a65-5855e7067288%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/f2c58590-8b35-4a62-a14c-3ec7babf2336%40continuum.io.
For more options, visit https://groups.google.com/a/continuum.io/d/optout.

Bryan,

I’m excited to hear that. Just an idea- it may be worth exploring how to integrate the front end with blaze in a smooth manner to facilitate rapid exploration and iteration of data and models. R’s piping is great, so maybe a more OO method chaining + patsy formula api for statsmodels would work.

Pandas has done similar with their recent addition of the “assign” method. (sorry if this has already been discussed).

datnamer

···

On Thursday, April 16, 2015 at 11:21:33 AM UTC-4, Bryan Van de ven wrote:

Yes there is actually an open PR that implements some of that, that will be in the 0.9 release next month. (In fact that’s exactly the sort of thing we want to “get right” before cranking on lots of new chart types)

Thanks,

Bryan

On Apr 16, 2015, at 10:18 AM, datnamer [email protected] wrote:

Sounds awesome! Would these chart types be compos-able and layer-able so we can mix and match as well?

On Thursday, April 16, 2015 at 10:36:37 AM UTC-4, Bryan Van de ven wrote:

Hi datnamer,

This is absolutely the plan. We have been working on clarifying the “charts” interface. This has involved some measure of experimentation to see what “fits” so we have not gone crazy adding new chart types while there is still this code churn. However, I we are now quite close to having a very stable charts API to build upon, and then we can start churning out all kinds of new chart types with confidence.

Thanks,

Bryan

On Apr 15, 2015, at 5:36 PM, datnamer [email protected] wrote:

Are there any plans for statistical glyphs for exploratory data analysis? I’m thinking ggplot stuff like smooth loess, kernel density plots, regressions etc


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/e92687c7-59c2-4863-9a65-5855e7067288%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/f2c58590-8b35-4a62-a14c-3ec7babf2336%40continuum.io.

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

Just FYI, charts already directly accept blaze expressions in many cases, but maybe you mean something else in addition?

Bryan

···

On Apr 16, 2015, at 5:03 PM, datnamer <[email protected]> wrote:

Bryan,

I'm excited to hear that. Just an idea- it may be worth exploring how to integrate the front end with blaze in a smooth manner to facilitate rapid exploration and iteration of data and models. R's piping is great, so maybe a more OO method chaining + patsy formula api for statsmodels would work.

Pandas has done similar with their recent addition of the "assign" method. (sorry if this has already been discussed).

datnamer

On Thursday, April 16, 2015 at 11:21:33 AM UTC-4, Bryan Van de ven wrote:
Yes there is actually an open PR that implements some of that, that will be in the 0.9 release next month. (In fact that's exactly the sort of thing we want to "get right" before cranking on lots of new chart types)

Thanks,

Bryan

> On Apr 16, 2015, at 10:18 AM, datnamer <[email protected]> wrote:
>
> Sounds awesome! Would these chart types be compos-able and layer-able so we can mix and match as well?
>
> On Thursday, April 16, 2015 at 10:36:37 AM UTC-4, Bryan Van de ven wrote:
> Hi datnamer,
>
> This is absolutely the plan. We have been working on clarifying the "charts" interface. This has involved some measure of experimentation to see what "fits" so we have not gone crazy adding new chart types while there is still this code churn. However, I we are now quite close to having a very stable charts API to build upon, and then we can start churning out all kinds of new chart types with confidence.
>
> Thanks,
>
> Bryan
>
> > On Apr 15, 2015, at 5:36 PM, datnamer <[email protected]> wrote:
> >
> > Are there any plans for statistical glyphs for exploratory data analysis? I'm thinking ggplot stuff like smooth loess, kernel density plots, regressions etc
> >
> > --
> > 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/e92687c7-59c2-4863-9a65-5855e7067288%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/f2c58590-8b35-4a62-a14c-3ec7babf2336%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/60dea255-8808-467d-9522-2331d3ff3f39%40continuum.io.
For more options, visit https://groups.google.com/a/continuum.io/d/optout.

Well it would be cool to specify models and plots with patsy ‘y~x’ formula type syntax. Just a thought , but probably too magical.

···

On Friday, April 17, 2015 at 8:53:58 AM UTC-4, Bryan Van de ven wrote:

Just FYI, charts already directly accept blaze expressions in many cases, but maybe you mean something else in addition?

Bryan

On Apr 16, 2015, at 5:03 PM, datnamer [email protected] wrote:

Bryan,

I’m excited to hear that. Just an idea- it may be worth exploring how to integrate the front end with blaze in a smooth manner to facilitate rapid exploration and iteration of data and models. R’s piping is great, so maybe a more OO method chaining + patsy formula api for statsmodels would work.

Pandas has done similar with their recent addition of the “assign” method. (sorry if this has already been discussed).

datnamer

On Thursday, April 16, 2015 at 11:21:33 AM UTC-4, Bryan Van de ven wrote:

Yes there is actually an open PR that implements some of that, that will be in the 0.9 release next month. (In fact that’s exactly the sort of thing we want to “get right” before cranking on lots of new chart types)

Thanks,

Bryan

On Apr 16, 2015, at 10:18 AM, datnamer [email protected] wrote:

Sounds awesome! Would these chart types be compos-able and layer-able so we can mix and match as well?

On Thursday, April 16, 2015 at 10:36:37 AM UTC-4, Bryan Van de ven wrote:
Hi datnamer,

This is absolutely the plan. We have been working on clarifying the “charts” interface. This has involved some measure of experimentation to see what “fits” so we have not gone crazy adding new chart types while there is still this code churn. However, I we are now quite close to having a very stable charts API to build upon, and then we can start churning out all kinds of new chart types with confidence.

Thanks,

Bryan

On Apr 15, 2015, at 5:36 PM, datnamer [email protected] wrote:

Are there any plans for statistical glyphs for exploratory data analysis? I’m thinking ggplot stuff like smooth loess, kernel density plots, regressions etc


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/e92687c7-59c2-4863-9a65-5855e7067288%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/f2c58590-8b35-4a62-a14c-3ec7babf2336%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/60dea255-8808-467d-9522-2331d3ff3f39%40continuum.io.

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