CODEWITHSUNDEEP
pythonpandasdataframebokehspectrogram
user1472672
user1472672

Reputation: 365

Drawing a heatmap in BokeH displays empty graph

I am trying to draw a heatmap(spectrogram) in bokeh but when the heatmap displays it is empty..

This is the code which has some simply sample data, but this would be extended to fetch a large dataset via json.

from math import pi
import pandas as pd

from bokeh.io import show
from bokeh.models import LinearColorMapper, BasicTicker, PrintfTickFormatter, ColorBar
from bokeh.plotting import figure

# initialise data of lists. 
data = {'epoch':[63745131000000, 63745131000000, 63745131100000,63745131100000], 'energy':[1.06811, 1.22078, 1.59495, 1.82245],'value':[3981.9308143034305, 2868.5202872178324, 1330.887696894385, 745.6847248644897]} 

# Creates pandas DataFrame. 
df = pd.DataFrame(data) 

# print the data 
print(df)

# this is the colormap from the original NYTimes plot
colors = ['#00007F', '#0000ff', '#007FFF', '#00ffff', '#7FFF7F', '#ffff00', '#FF7F00', '#ff0000', '#7F0000']
mapper = LinearColorMapper(palette=colors, low=df.value.min(), high=df.value.max())

TOOLS = "hover,save,pan,box_zoom,reset,wheel_zoom"

epochs = list(df.epoch.drop_duplicates())
print(epochs)
energies = list(df.energy.drop_duplicates())
print(energies)


p = figure(title="My Plot", 
           x_axis_location="below", 
           tools=TOOLS, toolbar_location='below',
           tooltips=[('epoch', '@epoch'), ('energy', '@energy'), ('value', '@value')])

p.xaxis.ticker = epochs
p.yaxis.ticker = energies

p.grid.grid_line_color = None
p.axis.axis_line_color = None
p.axis.major_tick_line_color = None
p.axis.major_label_text_font_size = "5pt"
p.axis.major_label_standoff = 0
p.xaxis.major_label_orientation = pi / 3

p.rect(x="epoch", y="energy", width=1, height=1,
       source=df,
       fill_color={'field': 'value', 'transform': mapper},
       line_color=None)

color_bar = ColorBar(color_mapper=mapper, major_label_text_font_size="5pt",
                     ticker=BasicTicker(desired_num_ticks=len(colors)),
                     label_standoff=6, border_line_color=None, location=(0, 0))
p.add_layout(color_bar, 'right')

show(p)

The frame which is output looks correct:

            epoch   energy        value
0  63745131000000  1.06811  3981.930814
1  63745131000000  1.22078  2868.520287
2  63745131100000  1.59495  1330.887697
3  63745131100000  1.82245   745.684725

and the ranges for the x and y look ok as well:

[63745131000000, 63745131100000]
[1.06811, 1.22078, 1.59495, 1.82245]

But the image that appears has no points plotted: enter image description here

I should mention, if I simply change the second epoch to one after e.g)

'epoch':[63745131000000, 63745131000000, 63745131000001,63745131000001]

Then the chart seems to be displayed correctly:

enter image description here

Grateful for any help. Thanks

Upvotes: 0

Views: 379

Answers (1)

Marc
Marc

Reputation: 1629

The reason why there is no information showing up is because at the edge bokeh apparently does not think your part of the data is something that deserves a color.

What you should change is the limits in you mapper:

mapper = LinearColorMapper(palette=colors, low=df.value.min()-1, high=df.value.max()+1) #  this will make sure your data is inside the mapping

Also your width is defined in the figure a being 1. When your epochs are differing with a million you will still see almost nothing when you are plotting this so change 

p.rect(x="epoch", y="energy", width=100000, height=1,  #  here width is set to an adequate level. 
       source=df,
       fill_color={'field': 'value', 'transform': mapper},
       line_color=None)

Upvotes: 2

Related Questions