# Copyright (c) 2017-2018 Spotify AB
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import chartify
import pandas as pd

def print_public_methods(obj):
    print('Methods:')
    print('\n'.join([x for x in dir(obj) if not x.startswith('_')]))

Chart object¶

Run the cell below to instantiate a chart and assign to a variable

In [ ]:

ch = chartify.Chart()

Use .show() to render the chart.

In [ ]:

ch.show()

Note that the chart is blank at this point.
The default labels provide directions for how to override their values.

Adding chart labels¶

Your turn: Add labels to the following chart. Look at the default values for instruction.
Title
Subtitle
Source
X label
Y label

In [ ]:

ch = chartify.Chart()
# Add code here to overwrite the labels


ch.show()

Getting help¶

From within a jupyter notebook you can see the available attributes of the chart object by pressing "tab"
Select the space just after the "." character below and hit tab.

In [ ]:

ch = chartify.Chart()
ch.

You can also use "?" to pull up documentation for objects and methods.
Run the cell below to pull up the chartify.Chart documentation

In [ ]:

chartify.Chart?

This can also be accomplished by pressing "shift + tab".
Press "shift + tab" twice to see the expanded documentation.
Try it with the next cell.

In [ ]:

chartify.Chart

Callouts¶

The chart object has a callout object (ch.callout) that contains methods for adding callouts to the chart.
Callouts can be used to add text, lines, or shaded areas to annotate parts of your chart.
Your Turn: Fill in the code below to add a text callout that says "hi" at coordinate (10, 10)
Look up the documentation for ch.callout.text if you need help

In [ ]:

ch = chartify.Chart()
ch.callout.text(# Fill in the code here)
    
ch.show()

Use tab below to see what callouts are available.

In [ ]:

ch.callout.

You should see this list of callouts:

In [ ]:

# List of available callouts:
print_public_methods(ch.callout)

Axes¶

The axes object contains methods for setting or getting axis properties.

In [ ]:

# Avaiable axes methods:
print_public_methods(ch.axes)

Your turn: modify the chart below so the xaxis range goes from 0 to 100

In [ ]:

ch = chartify.Chart()
ch.callout.text('hi', 10, 10)
# Add code here to modify the xrange to (0, 100)
ch.show()

Method chaining¶

Chart methods can be chained by wrapping the statments in parentheses. See the example below:

In [ ]:

(chartify.Chart(blank_labels=True)
 .callout.text('hi', 10, 10)
 .axes.set_xaxis_range(0, 100)
 .show()
)

Plotting¶

Input data format¶

Chartify expects the input data to be:

Tidy (Each variable has its own column, each row corresponds to an observation)
In the columns of a Pandas DataFrame.

Below we'll explore some examples of valid and invalid input data

Run this cell to generate an example dataset

In [ ]:

data = chartify.examples.example_data()
data.head()

Pivoted data: INVALID¶

Pivoted data is not Tidy (note the country dimension has an observation in each column)

In [ ]:

pivoted_data = pd.pivot_table(data, columns='country', values='quantity', index='fruit', aggfunc='sum')
pivoted_data

Melting pivoted data: VALID¶

You can use pandas.melt to convert pivoted data into the tidy data format.
The output of SQL queries with groupby produces output in tidy format.

In [ ]:

value_columns = pivoted_data.columns

melted_data = pd.melt(pivoted_data.reset_index(), # Need to reset the index to put "fruit" into a column.
                      id_vars='fruit',
                      value_vars=value_columns)
melted_data.head()

Pandas series: INVALID¶

Data in a pandas Series must be converted to a DataFrame for use with Chartify.

In [ ]:

data.groupby(['country'])['quantity'].sum()

Pandas index: INVALID¶

The output below is a pandas DataFrame, but the country dimension is in the Index.

In [ ]:

data.groupby(['country'])[['quantity']].sum()

Pandas DataFrame: VALID¶

The code below produces a valid pandas DataFrame for use with Chartify.
Notice how the country dimension is now in a column.

In [ ]:

chart_data = data.groupby(['country'])['quantity'].sum().reset_index()
chart_data

Axis types¶

Specify the x_axis_type and y_axis_type parameters when instantiating the chart object.
Both are set to linear by default.
Look at the chart object documentation to see the list of available options for x_axis_type and y_axis_type

In [ ]:

chartify.Chart?

The Chart axis types influence the plots that are available
Look at how the plot methods change based on the axis types:

In [ ]:

ch = chartify.Chart(x_axis_type='datetime',
                    y_axis_type='linear')
# List of available callouts:
print_public_methods(ch.plot)

In [ ]:

ch = chartify.Chart(x_axis_type='categorical',
                    y_axis_type='linear')
# List of available plots:
print_public_methods(ch.plot)

Your turn: Create a chart with 'density' y and 'linear' x axis types. What type of plots are available?

In [ ]:

ch = chartify.Chart(# Your code goes here)

Vertical Bar plot¶

Your turn: Create a bar plot based on the dataframe below.

In [ ]:

bar_data = (data.groupby('country')[['quantity']].sum()
            .reset_index()
           )
bar_data

In [ ]:

# Implement the bar plot here.
# Set the appropriate x_axis_type otherwise the bar method won't be available.
# Look at the bar documentation to figure out how to pass in the parameters.
# If you get stuck move on to the next section for hints.
ch = chartify.Chart(# Your code goes here)

Examples¶

Chartify includes many examples. They're a good starting point if you're trying to create a chart that you're unfamiliar with.

In [ ]:

# List of available examples
print_public_methods(chartify.examples)

Run the appropriate method to see examples and the corresponding code that generates the example.

In [ ]:

chartify.examples.plot_bar()

Bar plot - Horizontal vs. Vertical¶

Copy your bar plot here, but make it horizantal instead of vertical. Look to the example above if you get stuck.

In [ ]:

Grouped bar plot¶

Your Turn: Create a grouped bar plot with the data below.

In [ ]:

grouped_bar_data = (data.groupby(['country', 'fruit'])[['quantity']].sum()
            .reset_index()
           )
grouped_bar_data

In [ ]:

# Implement the grouped bar plot here.
# Look at the example for help if you get stuck.

show('html') vs. show('png')¶

Chartify charts can be rendered as either "HTML" or "PNG" (HTML is the default)
HTML output:
- Is faster and good for iteration.
- Can be saved as an image by screenshotting, or by clicking the icon on the top right.
- Will NOT show up in Jupyter notebooks when uploaded to GHE.
PNG output:
- Is slower and better for the finished product.
- Can be copy and pasted directly from the jupyter notebook (Right click on the image)
- Will show up in Jupyter notebooks when uploaded to GHE.
See the difference by running the two cells below

In [ ]:

(chartify.Chart(blank_labels=True)
.set_title("HTML output")
.set_subtitle("Faster, but will not show up in GHE")
.show()
)

In [ ]:

(chartify.Chart(blank_labels=True)
.set_title("PNG output")
.set_subtitle("Slower, but will show up in GHE. Right click to copy + paste.")
.show('png')
)

Color palette types¶

Chartify includes 4 different color palette types: categorical, accent, sequential, diverging.
Note the differences in the examples below

In [ ]:

chartify.examples.style_color_palette_categorical()

In [ ]:

chartify.examples.style_color_palette_accent()

In [ ]:

chartify.examples.style_color_palette_diverging()

In [ ]:

chartify.examples.style_color_palette_sequential()

Color palettes¶

Chartify includes a set of pre-defined color palettes:

In [ ]:

chartify.color_palettes

Use .show() to see the colors associated with each:

In [ ]:

chartify.color_palettes.show()

Assign the color palettes with .set_color_palette

In [ ]:

ch = chartify.Chart(x_axis_type='categorical',
                    blank_labels=True)
ch.style.set_color_palette('categorical', 'Dark2')
ch.plot.bar(data_frame=grouped_bar_data,
            categorical_columns=['fruit', 'country'],
            numeric_column='quantity',
            color_column='fruit')
ch.show()

Color palette objects include methods for manipulation. See the examples below:

In [ ]:

dark2 = chartify.color_palettes['Dark2']
dark2.show()

Sort

In [ ]:

sorted_dark2 = dark2.sort_by_hue()
sorted_dark2.show()

Expand

In [ ]:

dark2.expand_palette(20).show()

Shift

In [ ]:

shifted_dark2 = dark2.shift_palette('white', percent=20)
shifted_dark2.show()

In [ ]:

- Assign the shifted color palette to a chart:

In [ ]:

ch = chartify.Chart(x_axis_type='categorical',
                    blank_labels=True)
ch.style.set_color_palette('categorical', shifted_dark2)
ch.plot.bar(data_frame=grouped_bar_data,
            categorical_columns=['fruit', 'country'],
            numeric_column='quantity',
            color_column='fruit')
ch.show()

Layouts¶

Chartify layouts are tailored toward use in slides.
Notice how the output changes for each of the slide layout options below:

In [ ]:

layout_options = ['slide_100%', 'slide_75%', 'slide_50%', 'slide_25%']
for option in layout_options:
    ch = chartify.Chart(layout=option, blank_labels=True, x_axis_type='categorical')
    ch.set_title('Layout: {}'.format(option))
    ch.plot.bar(data_frame=grouped_bar_data,
            categorical_columns=['fruit', 'country'],
            numeric_column='quantity',
            color_column='fruit')

    ch.show()

Advanced usage with Bokeh¶

Chartify is built on top of another visualization package called Bokeh
The example below shows how you can access the Bokeh figure from a Chartify chart object.

In [ ]:

ch = chartify.Chart(blank_labels=True, x_axis_type='categorical')
ch.plot.bar(data_frame=grouped_bar_data,
        categorical_columns=['fruit', 'country'],
        numeric_column='quantity',
        color_column='fruit')
ch.figure

The following example shows how you can modify attributes not exposed in Chartify by accessing the Bokeh figure. See Bokeh documentation for more details.

In [ ]:

ch.figure.xaxis.axis_label_text_font_size = '30pt'
ch.figure.xaxis.axis_label_text_color = 'red'
ch.figure.height = 400
ch.axes.set_xaxis_label('A large xaxis label')
ch.show()

In [ ]:

Table of Contents

Chart object¶

Adding chart labels¶

Getting help¶

Callouts¶

Axes¶

Method chaining¶

Plotting¶

Input data format¶

Pivoted data: INVALID¶

Melting pivoted data: VALID¶

Pandas series: INVALID¶

Pandas index: INVALID¶

Pandas DataFrame: VALID¶

Axis types¶

Vertical Bar plot¶

Examples¶

Bar plot - Horizontal vs. Vertical¶

Grouped bar plot¶

show('html') vs. show('png')¶

Color palette types¶

Color palettes¶

Layouts¶

Advanced usage with Bokeh¶