CODEWITHSUNDEEP
pythonpandasbokeh
gtay
gtay

Reputation: 31

Making a slider that iterates over academic terms in a Bokeh Bar Chart

I'm making histograms of grades historically by term.

I want to make an interactive Bokeh bar chart with a slider that can cycle through the terms.

I have the bar chart working on a single term but when I try to add additional terms I can't get the bar chart to select a single term and then do the updating and slight through it.

I really just need to have some help getting the groupby object to select just one term. 

import os
from bokeh.plotting import figure, ColumnDataSource
from bokeh.io import curdoc, output_file, show
from bokeh.layouts import widgetbox
from bokeh.models import Slider
from bokeh.transform import jitter
from bokeh.palettes import viridis
from bokeh.transform import factor_cmap

Importing necessary modules. 

input_file = 'Alltime_grades.csv'

output_file('Grades.html')
df = pd.read_csv(input_file)
group = df.groupby(['Term','Grade'])

Here the is the input file import code. The input file has 3 columns - "Term", "Grade", and "Grade_Count".

So say Spring 2019 - A - 5000

I got this to work on a single term by not grouping by term even tho the field was still there. 

Grades = ['A',  'A-',   'B+',   'B',    'B-',   'C+',   'C',    'C-',   'D+',   'D',    'D-',   'F']

Above, I made a list of grades so that they display in the correct order

source = ColumnDataSource(group)

grade_cmap = factor_cmap('Grade',palette=viridis(22) , factors=Grades)
###
p = figure(plot_height=500, plot_width=700, title='Grades Over Time', toolbar_location=None, tools="", x_range = Grades)
###
p.vbar(x='Grade', top='grade_count_max', width=.75, bottom=0, source=source, line_color=grade_cmap, fill_color=grade_cmap)


p.y_range.start = 0
p.xaxis.axis_label = 'Grade'
p.yaxis.axis_label = 'Count of Grade'
curdoc().add_root(p)



show(p)

Above is the code I used to make the single bar chart to display. 

source = ColumnDataSource(data={
        'x' : group.loc['Spring 2019'].Grade,
        'top' :  group.loc['Spring 2019'].grade_count_max)})

For the multiple terms I tried this as the Column Data Source.

So for now I am getting can't use 'loc' method on Group by.

I need to have some way to make it select a single term so that I can them make an update function that follows the slider and updates the terms.

Beyond that I'm not even sure that I can cycle through non-numeric values on a slider but a good first step would be to be able to slice the group by at all.

Thanks for any help you can provide.

Upvotes: 0

Views: 116

Answers (2)

gtay
gtay

Reputation: 31

I finally figured this out by using the .get_group method on the GroupBy Object!

Upvotes: 0

Geomario
Geomario

Reputation: 212

I think you should describe your data while is not clear what is the main problem. This is a problem with pandas and not with bokeh, however, I ran with a similar problem but I cannot understand your data problem. The way I solved (in my case) was to use the pandas function Grouper and then I applied the mean. The ColumnDataSource for instance, allowed me to use the .loc function.

Upvotes: 0

Related Questions