Overview

Plotly is a Python library for creating interactive data visualizations. It builds on the Plotly JavaScript library which uses D3.

To install this package, enter pip install plotly.

To use Plotly inside a JupyterLab, assuming that is already installed, enter:

pip install "ipywidgets>=7.5"
jupyter labextension install jupyterlab-plotly@4.11.0
jupyter labextension install @jupyter-widgets/jupyterlab-manager plotlywidget@4.11.0

Command-line Example

This example generates a simple bar chart.

import plotly.graph_objects as go

data = [2, 3, 1]

# Create a bar chart using these values.
fig = go.Figure(data=go.Bar(y=data))

# Generate HTML for this bar chart, write it to a file,
# and open the file in the default web browser.
# The HTML consists of:
# - a script tag that configures Plotly
# - a script tag contains the Plotly library implementation
# - a script that that uses Plotly to draw the chart
# - a mostly empty svg element where the plot is created
fig.write_html('bar-chart.html', auto_open=True)

JupyterLab Example

This generates the same bar chart, but inside a Jupyter Notebook using JupyterLab. Enter jupyter lab to start it. Create a new Python notebook and enter the following:

import plotly.graph_objects as go

data = [2, 3, 1]
fig = go.Figure(data=go.Bar(y=data))
fig.show()

Note: This doesn't render a plot when run in JupyterLab launched from Anaconda Navigator!

Provided Datasets

The following code can be entered in a Jupyter Notebook to render a line chart that shows stock prices. It uses a dataset provided by Plotly in plotly.express.data. These include:

Dataset Name	Description
carshare	availability of car-sharing services in Montreal over a month-long period
election	voting results for a district in the 2013 Montreal mayoral election
election_geojson	each feature represents an electoral district in the 2013 Montreal mayoral election
experiment	results of 100 simulated participants on three hypothetical experiments
gapminder	life expectancy, population, and GDP for countries in a given year
iris	flower data
medals_long	Olympic medals for short track speed skating
medals_wide	variation on previous data
stocks	closing prices from six tech stocks in 2018 and 2019
tips	tips on restaurant bills
wind	wind intensity in a cardinal direction and its frequency

Line Chart

import plotly.express as px
import plotly.graph_objects as go

df = px.data.stocks()
px.line(
    df,
    x='date', y=['AAPL', 'GOOG'],
    labels={'x': 'Date', 'y': 'Price'},
    title="Apple vs. Google")

Hover over a data point to see a popup that shows its name and x/y values.

fig = go.Figure()
fig.add_trace(go.Scatter(x=df.date, y=df.AAPL, mode='lines', name='Apple'))
fig.add_trace(go.Scatter(x=df.date, y=df.AMZN, mode='lines+markers', name='Amazon'))
fig.add_trace(go.Scatter(
    x=df.date, y=df.GOOG, mode='lines+markers', name='Google',
    line=dict(color='firebrick', width=2, dash='dashdot')))

TODO: Add more detail!