Introduction to hvPlot

BackGround: Why hvplot?

Holoviz eco system

The Holoviz ecosystem is a suite of compatible tools designed to make data visualization in Python easier and more powerful. It supports every stage of data analysis, from importing and cleaning data, exploring, hypothesis testing, generating accurate and compelling diagrams, to sharing and deploying live applications.

The main features are as follows:

• Multiple tools work together to smoothly visualize data and create interactive dashboards.
• All tools are web browser compatible, providing the same usability in Jupyter notebooks and standalone applications.
• Easy to use with a high-level API, and allows for detailed customization as needed.
• Supports high-speed rendering of large-scale data and visualization of geographic information.

Representative components include the following:

HoloViews:

Separate data and its visual representation to easily create complex interactive diagrams

Panel:

Build interactive dashboards and web applications using only Python

hvPlot:

Easily create interactive plots from data structures such as pandas and xarray

Datashader:

Fast rendering of large datasets with millions of points or more

GeoViews:

Extend the visualization of geospatial data

Thus, Holoviz is built on top of the Python scientific and technological stack and is an ecosystem that comprehensively supports data visualization.

hvplot

hvplot is a high-level API developed as part of the Holoviz ecosystem (Holoviews, Panel, Datashader, etc.) that allows you to visualize pandas/xarray/Dask data directly.

Features of hvplot

• Plot pandas dataframes with matplotlib in the same way as df.plot()
• Automatically utilize Holoviews' powerful interaction features
• Switch between multiple backends such as Bokeh / Matplotlib / Plotly
• High-speed rendering of large datasets (millions of points) in conjunction with Datashader
• Interactive operation in Jupyter / Colab / VS Code

In particular, it is strong in meeting the need for "interactive visualization without increasing the amount of code."

Basic Structure of hvplot

hvplot operates as an extension of pandas.

import hvplot.pandas

This alone adds the .hvplot method to pandas DataFrames/Series. By default, Bokeh is used as the backend, but you can also specify matplotlib.

Basic Examples

1. !pip install hvplot
   import pandas as pd
   import matplotlib.pyplot as plt
   # Sample DataFrame
   data = {'Year': [2017, 2018, 2019, 2020, 2021],
           'Sales': [250, 300, 400, 350, 500]}
   df = pd.DataFrame(data)
   # Plotting a line graph
   df.plot(x='Year', y='Sales', kind='line')
   plt.xlabel('Year')
   plt.ylabel('Sales')
   plt.title('Yearly Sales')
   plt.show()
   

   

2. import hvplot.pandas
   #hvplot.extension('bokeh')
   plot = df.hvplot(x='Year', y='Sales', kind='line')
   plot.opts(xlabel='Year')
   plot.opts(ylabel='Sales')
   plot.opts(title='Yearly Sales')

   

3. hvplot.extension('matplotlib')
   plot2 = df.hvplot(x='Year', y='Sales',kind='line')
   plot2.opts(xlabel='Year')
   plot2.opts(ylabel='Sales')
   plot2.opts(title='Yearly Sales')

   

Strengths of hvplot: Interactivity

Because hvplot uses Holoviews internally, zoom, pan, hover, and toolbars are automatically included.

Without additional code, the following operations are possible:

• Select a range with the mouse and zoom
• Check values ​​on hover
• Save as PNG
• Reset

Plot

When drawing a graph with hvplot, specify the graph type as a string or as a method in the hbplot kind option.

line

Line graph

scatter

Scatter plot

bar

Bar graph

barh

Horizontal bar graph

hist

Histogram

box

Box plot

heatmap

Heatmap

* Pie charts are not officially supported.

You can adjust axis labels, colors, fonts, etc. using .opts.

Basic Visualization Examples

Line Graph

import pandas as pd
import hvplot.pandas
df=pd.DataFrame({'x':[1,2,3],'y':[4,5,2]})
df.hvplot.line(x='x',y='y')

Bar Graph

df=pd.DataFrame({'section': ['Div-A', 'Div-B', 'Div-c', 'Div-D', 'Div-E'],'students': [23,17,35,29,12]})
df.hvplot.bar(x='section',y='students')

Histgram

df=pd.DataFrame({'data':[12,80,30,65,60,73,55,54]})
plot=df.hvplot.hist(y='data', bins = 'auto')
plot.opts(title="HISTOGRAM")
plot.opts(xlabel='MARKS')
plot.opts(ylabel='STUDENTS')
plot

Scatter

girls_grades = [83, 90, 71, 80, 100, 87, 92, 99, 40, 34]
boys_grades = [70, 39, 59, 43, 100, 45, 74, 95, 87, 30]
grades_range = [10, 20, 30, 40, 50, 60, 70, 80, 90, 100]
df=pd.DataFrame({'girls_grades':girls_grades,'boys_grades':boys_grades,'grades_range':grades_range})
plot1=df.hvplot.scatter(x='grades_range', y='girls_grades', color='r')
plot2=df.hvplot.scatter(x='grades_range', y='boys_grades', color='b')
plot=plot1*plot2
plot.opts(xlabel='Grades Range')
plot.opts(ylabel='Grades Scored')
plot.opts(title='scatter plot')
plot

Conclusions

• hvplot is a high-level visualization library that integrates seamlessly with pandas.
• Powerful interactivity is built in by default.