Combine bokeh serve examples

Hi,

I’m trying to implement a visualisation tool which incorporates a few of the features across the Server App examples - http://bokeh.pydata.org/en/latest/docs/gallery.html

Essentially, I want a scatter plot which has changable X, Y, colour and size variables (like in the Crossfilter example). I would also like to incorporate the histograms (As in selected histogram example), and also the ability to filter data further (i.e like in the movies example, maybe also having a linked data table too.)

My problem is that the source codes are pretty different, and I’m not sure which is the best start point? Which code base should I start with and modify?

For example, crossfilter does the plotting in a function - create_figure() - and the update function simply consists of;

def update(attr, old, new):

layout.children[1] = create_figure()

``

Whereas the histogram code does not plot in a function, but updates the data_source generally. The movies example modifies the data source too.

I’ve tried to incorporate the histograms functionality by adding the histogram sections to the create_figure() plot of crossfilter, which does work, but I’m not sure where the update code should go?

Any pointers or more examples of combinations of the above would great!

Thanks