Archived discussions · Captured October 28, 2026

Map tile and x values limits

odadounDec 16, 2021 #1

Hi,

I have a bokeh figure which use x_axis_type=“mercator” and y_axis_type=“mercator”.
When I don’t use tile I obtain this map

As you can see the figure is encapsulated by the xmin,ymin, xmax and ymax values of my map.
Now, and with the same map, if I add a tile, esri for instance, I get this:

As you can see some value seems to be added to xmin and xmax.
Why this happen and is there a solution to remove it ?
By advance thank you.
cheers
Olivier
PS:By the way Bokeh gives very beautiful map

⌁ Link to post

ghomemDec 16, 2021 #2

Hi,

I don’t know if this is related. I use “None” as the tile provider and the map works fine without any distortion.

Cheers,

⌁ Link to post

ghomemDec 16, 2021 #3

In case you want to borrow some code:

github.com

ghomem/coviz/blob/master/main.py#L201

    
      
              if asource:
                  y_min, y_max = get_y_limits (asource, data_dates[0+DATE_IGNORE], data_dates[days-1])
                  range_delta = y_max * PLOT_RANGE_FACTOR
          
          
        # this thing alone prevents an interference from toggling the visibility of clines
                  # and the scale of the plots; comment this line and you will see :-)
                  # reference:
                  # https://discourse.bokeh.org/t/autoscaling-of-axis-range-with-streaming-multiline-plot-with-bokeh-server/1284/2?u=comperem
                  aplot.y_range=Range1d(y_min - range_delta , y_max + range_delta)
          
          
def make_map_plot( data ):
          
          
    hover_string = [ ('County', '@NAME_2'), ('Incidence', '@incidence'), ]
          
          
    # According to the docs the colormap parameter of the plot_bokeh function:
              # "Defines the colors to plot. Can be either a list of colors or the name of a Bokeh color palette"
          
          
    # because our original palette has the colors in the wrong direction
              # and because of this https://github.com/bokeh/bokeh/issues/7297
              # we can't just invert the colormap_range or we loose the legend on the color bar
              # so we we are forced to reverse the palette manually

⌁ Link to post

odadounDec 16, 2021 #4

thanks for the code. I didn’t know Pandas-Bokeh.
My problem is a boundary problem. I use total_bounds from my geopandas
to access the x and y limits of my map and use x_range in my bokeh figure.
But Bokeh figure seems to add an offset to this limit when tile is used …

⌁ Link to post

_jmDec 17, 2021 #5

It is a bit hard to speculate without a minimal reproducible example. The perspective I can add is that in the apps where I use an ESRI tile provider with bokeh I do the following to set up tile-with-overlay-plot graphics

do not explicitly set x_range or y_range, i.e. use the default behavior here
set match_aspect=True (in order to preserve the aspect ratio for my case where I am dynamically overlaying vehicle flight paths on the map)

With that in mind, is it the case that the extra buffer to the left/right of the map area of interest is due to wanting to preserve the correct aspect ratio? If you change the width/height parameters of your plot do these buffers follow suit?

⌁ Link to post

odadounDec 17, 2021 #6

Thanks for your reply.
You are right, Bokeh is trying to keep the aspect ratio and this take into account the plot_width, plot_height, x_range, y_range and the total_bounds calculation from geopandas should be in the loop somehow … I will try to find where this is implemented.
By the way plot_width=plot_height seems to give a better rendered.
thanks for you tips
cheers,
Olivier

⌁ Link to post

systemMar 17, 2022 #7

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.

⌁ Link to post

End of archived discussion · Replies are disabled