Bokeh + iPython for reporting? Makes any sense?

Hi,

I work as a management consultant. I often help customers prepare reports (for instance how sales are going per product/channel relative to budget). My tool of the trade was excel + powerpoint, which I have substituted with Tableau (a business analytics package).

I have recently discovered D3 and the python-pandas-bokeh stack. I understand that bokeh charts are not yet interactive like Tableau and D3 charts can be. In other words, I cannot build a bokeh chart and tie it to some kind of an interactive drop-down menu so the user can choose, for example, which the type of product the chart should display as I can do with D3.

My question is whether bokeh is developed also with business analytic users in mind (that demand this kind of interactivity and of “exploration”) or more for math/scientific applications and if it might already make sense to use iPython + Bokeh as a medium to share business reporting.

Thanks

Fabio

Hi Fabio,

It it currently possible to integrate BokehJS (the pure JavaScript library that is part of Bokeh) with any kinds of widgets or widget toolkits that you like. There are a few examples of this under the bokhejs directory. But you are correct that there is not really a way to tie these sorts of things together from the python side yet. That said, the use case you mention is absolutely a priority for Bokeh. Being able to create a plot that has widgets (sliders, dropdowns, linked data tables, etc.) attached to it, that can manipulate the plot (either through logic on the browser or full two way communication with a python server) is one of the reasons we created Bokeh in the first place! We want to be able to create dashboards and easily deployable "apps", exactly the way you describe. There are still a few more pieces to put into place, and we are also currently evaluating what the best options are for toolkits on the JS side. I would say to expect this functionality in the next couple of months. Peter may have some additional comments about the roadmap, as well.

Bryan

···

On Mar 12, 2014, at 7:48 AM, Fabio Annovazzi <[email protected]> wrote:

Hi,

I work as a management consultant. I often help customers prepare reports (for instance how sales are going per product/channel relative to budget). My tool of the trade was excel + powerpoint, which I have substituted with Tableau (a business analytics package).

I have recently discovered D3 and the python-pandas-bokeh stack. I understand that bokeh charts are not yet interactive like Tableau and D3 charts can be. In other words, I cannot build a bokeh chart and tie it to some kind of an interactive drop-down menu so the user can choose, for example, which the type of product the chart should display as I can do with D3.

My question is whether bokeh is developed also with business analytic users in mind (that demand this kind of interactivity and of "exploration") or more for math/scientific applications and if it might already make sense to use iPython + Bokeh as a medium to share business reporting.

Thanks

Fabio

--
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/4b4abe4b-695e-4678-b4ba-c83e45d4f9f1%40continuum.io.
For more options, visit https://groups.google.com/a/continuum.io/d/optout.

Thanks Bryan for the update on this. I certainly echo’s Fabio interest in these features, and it was also my understanding that Bokeh was started with those features as top priorities. As a data scientist working in the intersection of ML and visualization in industry, I see huge interest in Bokeh and its roadmap.

Like others, I looked into mpld3, vincent, plotly, ggplot-py with the hope of finding interactive rendering functionality and/or seamless integration between D3 and Python, but among all, Bokeh seems to have the highest GitHub repository activity and I believe is the most promising approach so far.

On this topic, I would be interested to see the breakdown of contributors on the repo (Continuum Analytics vs external contributors).

Keep up the good work!

Josh

···

On Wed, Mar 12, 2014 at 9:02 AM, Bryan Van de Ven [email protected] wrote:

Hi Fabio,

It it currently possible to integrate BokehJS (the pure JavaScript library that is part of Bokeh) with any kinds of widgets or widget toolkits that you like. There are a few examples of this under the bokhejs directory. But you are correct that there is not really a way to tie these sorts of things together from the python side yet. That said, the use case you mention is absolutely a priority for Bokeh. Being able to create a plot that has widgets (sliders, dropdowns, linked data tables, etc.) attached to it, that can manipulate the plot (either through logic on the browser or full two way communication with a python server) is one of the reasons we created Bokeh in the first place! We want to be able to create dashboards and easily deployable “apps”, exactly the way you describe. There are still a few more pieces to put into place, and we are also currently evaluating what the best options are for toolkits on the JS side. I would say to expect this functionality in the next couple of months. Peter may have some additional comments about the roadmap, as well.

Bryan

On Mar 12, 2014, at 7:48 AM, Fabio Annovazzi [email protected] wrote:

Hi,

I work as a management consultant. I often help customers prepare reports (for instance how sales are going per product/channel relative to budget). My tool of the trade was excel + powerpoint, which I have substituted with Tableau (a business analytics package).

I have recently discovered D3 and the python-pandas-bokeh stack. I understand that bokeh charts are not yet interactive like Tableau and D3 charts can be. In other words, I cannot build a bokeh chart and tie it to some kind of an interactive drop-down menu so the user can choose, for example, which the type of product the chart should display as I can do with D3.

My question is whether bokeh is developed also with business analytic users in mind (that demand this kind of interactivity and of “exploration”) or more for math/scientific applications and if it might already make sense to use iPython + Bokeh as a medium to share business reporting.

Thanks

Fabio

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/4b4abe4b-695e-4678-b4ba-c83e45d4f9f1%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/19C4CD9B-BB7E-47E5-857A-2FD8931DDC6C%40continuum.io.
For more options, visit https://groups.google.com/a/continuum.io/d/optout.

Very interesting! Thanks, It would be awesome.

Fabio

···

On Wednesday, March 12, 2014 12:48:12 PM UTC+1, Fabio Annovazzi wrote:

Hi,

I work as a management consultant. I often help customers prepare reports (for instance how sales are going per product/channel relative to budget). My tool of the trade was excel + powerpoint, which I have substituted with Tableau (a business analytics package).

I have recently discovered D3 and the python-pandas-bokeh stack. I understand that bokeh charts are not yet interactive like Tableau and D3 charts can be. In other words, I cannot build a bokeh chart and tie it to some kind of an interactive drop-down menu so the user can choose, for example, which the type of product the chart should display as I can do with D3.

My question is whether bokeh is developed also with business analytic users in mind (that demand this kind of interactivity and of “exploration”) or more for math/scientific applications and if it might already make sense to use iPython + Bokeh as a medium to share business reporting.

Thanks

Fabio

Bryan,

Are there plans to make it easier to create standard charts? For example, the new tutorial -- nice job, btw! -- shows how to create a bar chart. It's pretty clear, but it's not as easy as say matplotlib. The reason I'm asking is that I'm part of a team that's working on a project to make it easier for people who work in Excel and aren't programmers to switch to using Pandas for analyzing data. Bokeh produces beautiful graphs and has tremendous potential in the long run, but that's a fair amount to ask of non-techy users to create basic graphs.

Anders

Hi Josh,

I also do believe the approach we are taking, creating a full-featured, standalone JS library, that can be easily driven from Python (or Scala, or R or Julia, or Ruby, or Excel, or Matlab...) will provide the best set of building blocks for making powerful, flexible, easy-to-use tools for across the range of developers, domain experts, and casual analytics dashboard consumers. (We have a CEO who codes, but I don't think that's the norm!) Of course that leaves us with a big chunk of work to plough through on the front end, but we are making good good and steady progress now. :slight_smile:

I am excited to see how much the pace of development has increased over the last few releases, and also to see outside contributors start to join in. So, your question is a good one. Of the 40 or so contributors listed on GitHub, the top seven are all affiliated Continuum in one way or another. Most of the remaining contributors have added things like doc fixes and examples. However, we have had a few outside contributions of substantial size: python 3 support, datetime axis scales, and a bunch of tests. We'd love to see even more people get involved at this kind of level. Certainly, if there's anyone on the list or elsewhere that wants to get involved, I will bend over backward to make sure they get plugged in and get all their questions answered.

Bryan

···

On Mar 12, 2014, at 9:28 AM, Josh Wasserstein <[email protected]> wrote:

Thanks Bryan for the update on this. I certainly echo's Fabio interest in these features, and it was also my understanding that Bokeh was started with those features as top priorities. As a data scientist working in the intersection of ML and visualization in industry, I see huge interest in Bokeh and its roadmap.

Like others, I looked into mpld3, vincent, plotly, ggplot-py with the hope of finding interactive rendering functionality and/or seamless integration between D3 and Python, but among all, Bokeh seems to have the highest GitHub repository activity and I believe is the most promising approach so far.

On this topic, I would be interested to see the breakdown of contributors on the repo (Continuum Analytics vs external contributors).

Keep up the good work!

Josh

On Wed, Mar 12, 2014 at 9:02 AM, Bryan Van de Ven <[email protected]> wrote:
Hi Fabio,

It it currently possible to integrate BokehJS (the pure JavaScript library that is part of Bokeh) with any kinds of widgets or widget toolkits that you like. There are a few examples of this under the bokhejs directory. But you are correct that there is not really a way to tie these sorts of things together from the python side yet. That said, the use case you mention is absolutely a priority for Bokeh. Being able to create a plot that has widgets (sliders, dropdowns, linked data tables, etc.) attached to it, that can manipulate the plot (either through logic on the browser or full two way communication with a python server) is one of the reasons we created Bokeh in the first place! We want to be able to create dashboards and easily deployable "apps", exactly the way you describe. There are still a few more pieces to put into place, and we are also currently evaluating what the best options are for toolkits on the JS side. I would say to expect this functionality in the next couple of months. Peter may have some additional comments about the roadmap, as well.

Bryan

On Mar 12, 2014, at 7:48 AM, Fabio Annovazzi <[email protected]> wrote:

> Hi,
>
> I work as a management consultant. I often help customers prepare reports (for instance how sales are going per product/channel relative to budget). My tool of the trade was excel + powerpoint, which I have substituted with Tableau (a business analytics package).
>
> I have recently discovered D3 and the python-pandas-bokeh stack. I understand that bokeh charts are not yet interactive like Tableau and D3 charts can be. In other words, I cannot build a bokeh chart and tie it to some kind of an interactive drop-down menu so the user can choose, for example, which the type of product the chart should display as I can do with D3.
>
> My question is whether bokeh is developed also with business analytic users in mind (that demand this kind of interactivity and of "exploration") or more for math/scientific applications and if it might already make sense to use iPython + Bokeh as a medium to share business reporting.
>
> Thanks
>
> Fabio
>
> --
> 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/4b4abe4b-695e-4678-b4ba-c83e45d4f9f1%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/19C4CD9B-BB7E-47E5-857A-2FD8931DDC6C%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/CAD4ivxUU3%2Bhi%2BSsSpqvASPPnjuR%2Bv9m10vp%2BBkfGAUhOK1wi6g%40mail.gmail.com.
For more options, visit https://groups.google.com/a/continuum.io/d/optout.

Hi Anders,

Absolutely! Something like this very nearly made it into the 0.4.2 release yesterday, but I wanted to have a slightly wider discussion about the interface before just churning out code. There is actually a ticket already to discuss this bokeh.charts interface:

  https://github.com/ContinuumIO/bokeh/issues/393

Some obvious types of charts we plan to include:

* bars (stacked, grouped, percentage, superimposed)
* hbars (stacked, grouped, percentage, superimposed)
* areas (stacked, superimposed)
* histogram (single or superimposed)
* boxplot
* lines

But if you have specific ideas about types of charts, interfaces, etc. please feel free to comment on that ticket (or here in the group). I actually plan to work on this in the next two weeks, so it should be in GH on master in that time frame, and part of the next 0.5 release in early April.

I should mention one other feature that just occurred to us to add: a command like bokeh tool, ala "gplot" so that could do things like:

  $ cat foo.csv | bokeh hbar($1, $4, $5, $6) outfile.html

or for sending updating data to server output:

  $ tail -f mylog | bokeh scatter($time, $errors) --host <plot server host>

Syntax very much TBD, but hopefully you get the idea!

Bryan

···

On Mar 12, 2014, at 9:57 AM, [email protected] wrote:

Bryan,

Are there plans to make it easier to create standard charts? For example, the new tutorial -- nice job, btw! -- shows how to create a bar chart. It's pretty clear, but it's not as easy as say matplotlib. The reason I'm asking is that I'm part of a team that's working on a project to make it easier for people who work in Excel and aren't programmers to switch to using Pandas for analyzing data. Bokeh produces beautiful graphs and has tremendous potential in the long run, but that's a fair amount to ask of non-techy users to create basic graphs.

Anders

--
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/48c56e62-e6e9-420e-87dd-fdfdfeda8b88%40continuum.io.
For more options, visit https://groups.google.com/a/continuum.io/d/optout.

Hi Brian,

Great! I'd add pie charts -- yeah, I know we're not supposed to use them most of the time, but they do come in handy for some situations and most user expect them -- and bubble charts.

FYI, our goal is to not only make it easy for non-techies who're Excel users to switch to pandas but also to start building libraries on top of pandas and graphing & stats libraries to help smooth out the learning curve of doing data science. For example, we're hoping to figure out how to make it as easy to create a chart that follows best practices of data vizualization as to create the hot mess that most folks usually build. We're just getting started -- including learning pandas :slight_smile: -- but once we've got a little more experience I think the team will definitely want to get involved in Bokeh.

Thanks,
Anders

Great! I'd add pie charts -- yeah, I know we're not supposed to use them most of the time, but they do come in handy for some situations and most user expect them -- and bubble charts.

I am torn on this too, but it does probably make sense to include them. Maybe by default they can be covered with an annoying watermark that can only be turned of by setting a "pie_charts_are_bad_mmmkay=True" keyword argument explicitly. :slight_smile:

FYI, our goal is to not only make it easy for non-techies who're Excel users to switch to pandas but also to start building libraries on top of pandas and graphing & stats libraries to help smooth out the learning curve of doing data science. For example, we're hoping to figure out how to make it as easy to create a chart that follows best practices of data vizualization as to create the hot mess that most folks usually build. We're just getting started -- including learning pandas :slight_smile: -- but once we've got a little more experience I think the team will definitely want to get involved in Bokeh.

Closer Pandas integration is definitely on our Roadmap as well, so definitely let us know what your use-cases are and we will keep them in mind. We are also working on a higher level open-source tool "CDX" that would use bokeh and pandas and provide a workspace-like area with linked tables and charts in the browser, something akin to R Studio. It's in a pretty nascent state right at the moment but you can check it out now at

  https://github.com/ContinuumIO/cdx

Peter can definitely comment more about the future plans and directions for CDX but again we'd love to get as much input on use-cases as we can, as well as collaborate wherever we can on pushing tools forward. Definitely whenever you are ready to get involved, let us know and we will make sure you get all the information and support you need.

Bryan

···

On Mar 12, 2014, at 10:59 AM, [email protected] wrote:

Oooo, that looks interesting! We've been kicking around some ideas on softening/blurring the lines between interacting visually and using coding/recipies, so we might have some uses cases to contribute once we're a bit further along.

Thanks for being so interested in having more folks involved! We'll definitely be in touch.

Anders

Hey I just wanted to post an update on this. We have not forgotten the bokeh.charts interface, in fact we have had some very nice discussions about it. We'd really like to incorporate the ability to to faceting easily in this interface. So for instance if you have data with the columns: year, quarter, symbol, price then you could do:

  bar(data, [facet("year"), group("quarter")], ...)

Or something similar, and you would get a separate chart for each year, each chart with bars for every symbol/price, grouped by quarter. But, this will rely on having much better layout which is work that is ongoing now. I'd still like to get in a few simple charts into the upcoming 0.5 release, but the full interface may take until 0.6. If folks have any thoughts or ideas about this, please chime in!

Bryan

···

On Mar 12, 2014, at 11:33 AM, Bryan Van de Ven <[email protected]> wrote:

On Mar 12, 2014, at 10:59 AM, [email protected] wrote:

Great! I'd add pie charts -- yeah, I know we're not supposed to use them most of the time, but they do come in handy for some situations and most user expect them -- and bubble charts.

I am torn on this too, but it does probably make sense to include them. Maybe by default they can be covered with an annoying watermark that can only be turned of by setting a "pie_charts_are_bad_mmmkay=True" keyword argument explicitly. :slight_smile:

FYI, our goal is to not only make it easy for non-techies who're Excel users to switch to pandas but also to start building libraries on top of pandas and graphing & stats libraries to help smooth out the learning curve of doing data science. For example, we're hoping to figure out how to make it as easy to create a chart that follows best practices of data vizualization as to create the hot mess that most folks usually build. We're just getting started -- including learning pandas :slight_smile: -- but once we've got a little more experience I think the team will definitely want to get involved in Bokeh.

Closer Pandas integration is definitely on our Roadmap as well, so definitely let us know what your use-cases are and we will keep them in mind. We are also working on a higher level open-source tool "CDX" that would use bokeh and pandas and provide a workspace-like area with linked tables and charts in the browser, something akin to R Studio. It's in a pretty nascent state right at the moment but you can check it out now at

  https://github.com/ContinuumIO/cdx

Peter can definitely comment more about the future plans and directions for CDX but again we'd love to get as much input on use-cases as we can, as well as collaborate wherever we can on pushing tools forward. Definitely whenever you are ready to get involved, let us know and we will make sure you get all the information and support you need.

Bryan

HI guys, I just wanted to give an update on this thread mentioning some of the things that just landed in the 0.5 release:

  • bokeh.charts was finally started thanks to Damian Avila. It has currently Bar, Histogram, and Scatter charts, but a Bokxplot chart PR was worked up by a new contributor in just a few hours at the SciPy sprint. It will go in 0.5.1 this week and we look to add many new chart types in the next month

  • widgets are not part of Bokeh proper, and values from widgets can be sent back to python processes and the bokeh-server to update the plot, etc.

Now is a great time to try Bokeh for dash boarding and reporting, let us know if we can help!

Bryan

···

On Monday, March 24, 2014 11:23:20 AM UTC-4, Bryan Van de ven wrote:

Hey I just wanted to post an update on this. We have not forgotten the bokeh.charts interface, in fact we have had some very nice discussions about it. We’d really like to incorporate the ability to to faceting easily in this interface. So for instance if you have data with the columns: year, quarter, symbol, price then you could do:

    bar(data, [facet("year"), group("quarter")], ...)

Or something similar, and you would get a separate chart for each year, each chart with bars for every symbol/price, grouped by quarter. But, this will rely on having much better layout which is work that is ongoing now. I’d still like to get in a few simple charts into the upcoming 0.5 release, but the full interface may take until 0.6. If folks have any thoughts or ideas about this, please chime in!

* widgets are not part of Bokeh proper, and values from widgets can be
sent back to python processes and the bokeh-server to update the plot,
etc.

I think this is supposed to read "now part of Bokeh proper"

Hi Fabio and Python folks!

You might be interested in these resources for automatically generating email, PDF, or HTML reports with Plotly and IPython notebook:

https://plot.ly/ipython-notebooks/gallery/#report-generation

Hopefully helpful! Thanks!

···

On Wednesday, March 12, 2014 at 4:48:12 AM UTC-7, Fabio Annovazzi wrote:

Hi,

I work as a management consultant. I often help customers prepare reports (for instance how sales are going per product/channel relative to budget). My tool of the trade was excel + powerpoint, which I have substituted with Tableau (a business analytics package).

I have recently discovered D3 and the python-pandas-bokeh stack. I understand that bokeh charts are not yet interactive like Tableau and D3 charts can be. In other words, I cannot build a bokeh chart and tie it to some kind of an interactive drop-down menu so the user can choose, for example, which the type of product the chart should display as I can do with D3.

My question is whether bokeh is developed also with business analytic users in mind (that demand this kind of interactivity and of “exploration”) or more for math/scientific applications and if it might already make sense to use iPython + Bokeh as a medium to share business reporting.

Thanks

Fabio

Do you really suppose this sort of off-topic shilling is appropriate for an open source mailing list?

Bryan

···

On Jan 8, 2016, at 19:55, [email protected] wrote:

Hi Fabio and Python folks!

You might be interested in these resources for automatically generating email, PDF, or HTML reports with Plotly and IPython notebook:

https://plot.ly/ipython-notebooks/gallery/#report-generation

Hopefully helpful! Thanks!

On Wednesday, March 12, 2014 at 4:48:12 AM UTC-7, Fabio Annovazzi wrote:

Hi,

I work as a management consultant. I often help customers prepare reports (for instance how sales are going per product/channel relative to budget). My tool of the trade was excel + powerpoint, which I have substituted with Tableau (a business analytics package).

I have recently discovered D3 and the python-pandas-bokeh stack. I understand that bokeh charts are not yet interactive like Tableau and D3 charts can be. In other words, I cannot build a bokeh chart and tie it to some kind of an interactive drop-down menu so the user can choose, for example, which the type of product the chart should display as I can do with D3.

My question is whether bokeh is developed also with business analytic users in mind (that demand this kind of interactivity and of “exploration”) or more for math/scientific applications and if it might already make sense to use iPython + Bokeh as a medium to share business reporting.

Thanks

Fabio

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/5cb37b2f-6508-4974-a166-6f19fc4ce2f3%40continuum.io.

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

That said, thanks for bringing this year and a half old thread back to our attention. With the recent release of Bokeh 0.11, the answer to the OP really needs an update!

To start, the new "bokeh" command is quite flexible, in addition to being able to run

  bokeh serve myapp.py

to bea able to generate live (not static), interactive apps and dashboards, that can query data, perform model computations, or stream updates, such as the live hosted examples shown at

  http://demo.bokehplots.com

and source examples at

  https://github.com/bokeh/bokeh/tree/master/examples/app

The "bokeh" command line tool also has other modes. Including "bokeh html", which can be used to take the *exact same source input* and generate static HTML:

  bokeh html myapp.py

Fairly soon (there is a PR in flight) you'll be able to do things like

  bokeh svg myapp.py

to generate static outputs in other formats as well. I am also very interested in a command to turn Jupyter notebooks directly into apps or HTML output. The good news is that with the new architecture that was build with the help of Havoc Pennington, this will be fairly trivial to do. In fact, it just moved to the top of my list.

Thanks,

Bryan

···

On Jan 8, 2016, at 8:26 PM, Bryan Van de Ven <[email protected]> wrote:

Do you really suppose this sort of off-topic shilling is appropriate for an open source mailing list?

Bryan

On Jan 8, 2016, at 19:55, [email protected] wrote:

Hi Fabio and Python folks!

You might be interested in these resources for automatically generating email, PDF, or HTML reports with Plotly and IPython notebook:

https://plot.ly/ipython-notebooks/gallery/#report-generation

Hopefully helpful! Thanks!

On Wednesday, March 12, 2014 at 4:48:12 AM UTC-7, Fabio Annovazzi wrote:
Hi,

I work as a management consultant. I often help customers prepare reports (for instance how sales are going per product/channel relative to budget). My tool of the trade was excel + powerpoint, which I have substituted with Tableau (a business analytics package).

I have recently discovered D3 and the python-pandas-bokeh stack. I understand that bokeh charts are not yet interactive like Tableau and D3 charts can be. In other words, I cannot build a bokeh chart and tie it to some kind of an interactive drop-down menu so the user can choose, for example, which the type of product the chart should display as I can do with D3.

My question is whether bokeh is developed also with business analytic users in mind (that demand this kind of interactivity and of "exploration") or more for math/scientific applications and if it might already make sense to use iPython + Bokeh as a medium to share business reporting.

Thanks

Fabio

--
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/5cb37b2f-6508-4974-a166-6f19fc4ce2f3%40continuum.io.
For more options, visit https://groups.google.com/a/continuum.io/d/optout.

Thanks to this lovely reminder, there is now a PR to support creating Bokeh apps directly from Jupyter notebooks! You can check it out here:

  https://github.com/bokeh/bokeh/pull/3594

There's a few small tasks, but I have a 6:30 AM flight. I'll finish them on the plane, and you can expect to see this in an easily installable dev build early next week, and in a planned 0.11.1 release very soon as well.

Thanks,

Bryan

···

On Jan 8, 2016, at 8:54 PM, Bryan Van de Ven <[email protected]> wrote:

That said, thanks for bringing this year and a half old thread back to our attention. With the recent release of Bokeh 0.11, the answer to the OP really needs an update!

To start, the new "bokeh" command is quite flexible, in addition to being able to run

  bokeh serve myapp.py

to bea able to generate live (not static), interactive apps and dashboards, that can query data, perform model computations, or stream updates, such as the live hosted examples shown at

  http://demo.bokehplots.com

and source examples at

  https://github.com/bokeh/bokeh/tree/master/examples/app

The "bokeh" command line tool also has other modes. Including "bokeh html", which can be used to take the *exact same source input* and generate static HTML:

  bokeh html myapp.py

Fairly soon (there is a PR in flight) you'll be able to do things like

  bokeh svg myapp.py

to generate static outputs in other formats as well. I am also very interested in a command to turn Jupyter notebooks directly into apps or HTML output. The good news is that with the new architecture that was build with the help of Havoc Pennington, this will be fairly trivial to do. In fact, it just moved to the top of my list.

Thanks,

Bryan

On Jan 8, 2016, at 8:26 PM, Bryan Van de Ven <[email protected]> wrote:

Do you really suppose this sort of off-topic shilling is appropriate for an open source mailing list?

Bryan

On Jan 8, 2016, at 19:55, [email protected] wrote:

Hi Fabio and Python folks!

You might be interested in these resources for automatically generating email, PDF, or HTML reports with Plotly and IPython notebook:

https://plot.ly/ipython-notebooks/gallery/#report-generation

Hopefully helpful! Thanks!

On Wednesday, March 12, 2014 at 4:48:12 AM UTC-7, Fabio Annovazzi wrote:
Hi,

I work as a management consultant. I often help customers prepare reports (for instance how sales are going per product/channel relative to budget). My tool of the trade was excel + powerpoint, which I have substituted with Tableau (a business analytics package).

I have recently discovered D3 and the python-pandas-bokeh stack. I understand that bokeh charts are not yet interactive like Tableau and D3 charts can be. In other words, I cannot build a bokeh chart and tie it to some kind of an interactive drop-down menu so the user can choose, for example, which the type of product the chart should display as I can do with D3.

My question is whether bokeh is developed also with business analytic users in mind (that demand this kind of interactivity and of "exploration") or more for math/scientific applications and if it might already make sense to use iPython + Bokeh as a medium to share business reporting.

Thanks

Fabio

--
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/5cb37b2f-6508-4974-a166-6f19fc4ce2f3%40continuum.io.
For more options, visit https://groups.google.com/a/continuum.io/d/optout.