14- Plotting in Python

14- Plotting in Python

In this tutorial, you will learn how to plot different types of visualizations such as line plots, bar plots, pair plots, scatter plots, joint plots, pie charts, box plots, histograms, animated plots, and various types of categorical plots using libraries such as Matplotlib, Seaborn, and others.

First, download the datasets that we will be using in this tutorial: london_borough_profiles1.csv, myPub.csv, and sar_data.csv.

Importing libraries

import re
import numpy as np
import pandas as pd
import random
import matplotlib as mpl
import matplotlib.pyplot as plt
import matplotlib.cbook as cbook
import seaborn as sns
import pandas_datareader as pdr

from pylab import *
from mpl_toolkits.mplot3d import Axes3D
from matplotlib.colors import ListedColormap

Time plot (time series)

To plot the time variation of a single variable, use the following code:

# Load time series data at Github.
df = pd.read_csv('https://raw.githubusercontent.com/selva86/datasets/master/a10.csv', parse_dates=['date'], index_col='date')

# define 'tplot' function
def tplot(df, x, y, title="", xlabel='Date', ylabel='Value', dpi=300):
    plt.figure(figsize=(16,5), dpi=dpi)
    plt.plot(x, y, marker='o', markerfacecolor='blue')
    plt.gca().set(title=title, xlabel=xlabel, ylabel=ylabel)
    plt.show()

tplot(df, x=df.index, y=df.value, title='Anti-diabetic sales in Australia from 1992 to 2008.')

image.png

Seasonal plot of a time series

To plot a seasonal plot of a time series, use the following code:


# Load Data
df = pd.read_csv('https://raw.githubusercontent.com/selva86/datasets/master/a10.csv', parse_dates=['date'], index_col='date')
df.reset_index(inplace=True)

# Prepare data
df['year'] = [d.year for d in df.date]
df['month'] = [d.strftime('%b') for d in df.date]
years = df['year'].unique()

# Prepare Colors
np.random.seed(50)
mycolors = np.random.choice(list(mpl.colors.XKCD_COLORS.keys()), len(years), replace=False)

# Draw Plot
plt.figure(figsize=(15,10), dpi= 300)
for i, y in enumerate(years):
    if i > 0:        
        plt.plot('month', 'value', data=df.loc[df.year==y, :], color=mycolors[i], label=y)
        plt.text(df.loc[df.year==y, :].shape[0]-.9, df.loc[df.year==y, 'value'][-1:].values[0], y, fontsize=12, color=mycolors[i])

# Decoration
plt.gca().set(xlim=(-0.3, 11), ylim=(2, 30), ylabel='Drug Sales', xlabel='Month')
plt.yticks(fontsize=12, alpha=.7)
plt.subtitle("Seasonal plot of time series", fontsize=18)
plt.show()

image.png

Plotting two variables time series:


rs = np.random.RandomState(365) # create data
values = rs.randn(365, 2).cumsum(axis=0)
dates = pd.date_range("1 1 2021", periods=365, freq="D")
data = pd.DataFrame(values, dates, columns=["A", "B"])
data = data.rolling(7).mean()

sns.lineplot(data=data, palette="tab10", linewidth=2.5)

image.png

Annotated heatmaps


# Load the example flights dataset and convert to long-form
flights_long = sns.load_dataset("flights")
flights = flights_long.pivot("month", "year", "passengers")

# Draw a heatmap with the numeric values in each cell
f, ax = plt.subplots(figsize=(9, 6))
sns.heatmap(flights, annot=True, fmt="d", linewidths=.5, ax=ax)

image.png

Swarmplot

Swarmplot used to display distribution of attributes.


# Import csv file of data

df = pd.read_csv ('D:/Python/Python_for_Researchers/london_borough_profiles1.csv', encoding='unicode_escape')
df.head()


# Create dataframe from some columns

df = df[['In_Out','Inner/_Outer_London', 'Happiness_score_2011-14_(out_of_10)', 'Anxiety_score_2011-14_(out_of_10)','Employment_rate_(%)_(2015)'
,'People_aged_17+_with_diabetes_(%)']]


# Cleaning data by change some names of columns

df.rename(columns={'Inner/_Outer_London': 'in_out','Happiness_score_2011-14_(out_of_10)':'happiness', 'Anxiety_score_2011-14_(out_of_10)':'anxiety', 'Employment_rate_(%)_(2015)':'employment','People_aged_17+_with_diabetes_(%)':'diabetes' }, inplace=True)

# Create some different data frames

df = df.reindex(columns=['diabetes', 'In_Out','in_out', 'happiness', 'anxiety', 'employment'])
df1 = df[['diabetes', 'happiness', 'anxiety']]
df2 = df[['in_out', 'employment', 'happiness', 'anxiety']]

# Create swarmplot using seaborn library

sns.swarmplot(data=df1)
plt.gca().set(ylabel='Value', xlabel='Indices') # set x and y labels

image.png

Barplot

A bar chart or bar graph is used to present categorical data.

sns.barplot(data=df1)
plt.gca().set(ylabel='Value', xlabel='Indices')

image.png

Stacked Barplot

import matplotlib.pyplot as plt
labels = ['A', 'B', 'C', 'D', 'E']
men_av = [23, 25, 33, 30, 18]
women_av = [15, 22, 30, 10, 15]
std_m = [1, 2.5, 3, 1, 1.5]
std_w = [2, 4, 1.5, 2, 2.5]
width = 0.5       # the width of the bars: can also be len(x) sequence

fig, ba = plt.subplots()

ba.bar(labels, men_av, width, yerr=std_m, label='Men')
ba.bar(labels, women_av, width, yerr=std_w,
       label='Women')

ba.set_ylabel('Scores')
ba.set_title('Scores by group and gender')
ba.legend()

plt.show()

image.png

Pairplot

Pairplot is used to present the distribution of variables and relationships between variables.

sns.pairplot(data=df2, hue='in_out')

image.png

Scatterplot

A scatter plot shows the relationship between two variables.

sns.scatterplot(data=df2, x = 'employment',y= 'happiness', hue='in_out')
plt.legend(title="", loc=8)

image.png

3D Scatterplot



sns.set_style("whitegrid", {'axes.grid' : False})
fig = plt.figure()
ax = Axes3D(fig) # Method 1
#ax = fig.add_subplot(111, projection='3d') # Method 2

Create x, y, and z NumPy array data.

X = np.array([0, 5, 10, 15, 20, 22, 26, 24, 14, 30])
Y = np.array([0, 3, 6, 9, 12, 22, 24, 26, 30, 20])
Z = np.array([3, 5, 11, 10, 12, 4, 5, 17, 10, 13])

Get a colormap from the seaborn library.

cmap = ListedColormap(sns.color_palette("husl", 256).as_hex())
g = ax.scatter(X, Y, Z, c=X, s= 50, marker='o', cmap = cmap, alpha = 1)

Set x, y, and z labels

ax.set_xlabel('X Label')
ax.set_ylabel('Y Label')
ax.set_zlabel('Z Label')

Add a color bar that maps values to colors.

fig.colorbar( g, shrink=0.5, aspect=5)
plt.show()

image.png

Use varying marker colors and sizes for the scatter plot.

Load data from Pandas Datareader using the (^DJI stooq) function.

data = pdr.DataReader('^DJI', 'stooq')# Data of ^DJI stooq market
data
data = data[-365:] # get the most recent 365 days data
delta1 = np.diff(data.Close) / data.Close[:-1] # price of close day / price of close day before
volume = (15 * data.Volume[:-2] / data.Volume[0])**2

Set colors for 363 days from the seaborn library's color palette.

colors = sns.color_palette("Set3", 363)

Plotting to scatter plot:

fig, pl = plt.subplots()
pl.scatter(delta1[:-1], delta1[1:], color=colors, s=volume, alpha = 0.5)

Set x, y labels, and title:

pl.set_xlabel(r'Δi', fontsize=12)
pl.set_ylabel(r'Δi+1', fontsize=12)
pl.xaxis.label.set_color('midnightblue')
pl.yaxis.label.set_color('midnightblue')
pl.set_title('Scatter plot of ^DJI stooq with volume and price change')
pl.grid(True)

Set x, y limitations

pl.axis([-0.025, 0.025, -0.025, 0.025]) # xlim , ylim
fig.tight_layout()
plt.show()

image.png

Jointplot

In addition to showing the relationship between the dependent variable (Y) and independent variable (X), it displays the distribution of X and Y.

# Linear regression
x = (1,3,5, 2, 9, 11)
y = (2,4,6, 3, 8, 10)
sns.jointplot (x=x, y=y, data =df , kind = "reg")
plt.gca().set(ylabel='Y', xlabel='X')

image.png

Piechart


# create data
names='A', 'B', 'C', 'D',
values=[5, 15, 30, 50]

# create a pieplot
plt.pie(values, labels = names, labeldistance=1.15, shadow=True, startangle=90, autopct='%1.1f%%')# Label distance: gives the space between labels and the center of the pie
plt.show()

image.png

Boxplot

df = pd.read_csv (r'D:\Python\Python_for_Researchers\sar_data.csv', encoding='unicode_escape')
df.head()

sns.boxplot(data=df,palette=["m", "g"])
sns.despine(offset=10, trim=True)
plt.gca().set(ylabel='Value', xlabel='Sensor')

image.png

Histogram

It represents the distribution of numerical data.

bio = [-2, 1, 2, 4, 2, 5, 5, 5,6 , 7, 9, 7, 5, 10, 15] # create data

sns.set_style('darkgrid') # set grid style
his = sns.distplot(bio)
his.set_xlabel('Value', fontsize=12) # set x label
his.set_ylabel('Frequency', fontsize=12) # set y label

image.png

Animated plot in Python

import pandas as pd
import matplotlib.pyplot as plt
from PIL import Image
import io

# Read the data
data = pd.read_csv(r'd://myPub1.csv')

# Check the first 5 rows
data.head(5)

# Transform the categorical column (Open) into numerical values
data['Open'] = pd.Categorical(data['Open'])

# Initialize a list to store images
images = []

# For each year:
for i in data.Year.unique():
    # Turn interactive plotting off
    plt.ioff()

    # Initialize a figure
    fig = plt.figure(figsize=(10, 6))

    # Find the subset of the dataset for the current year
    subsetData = data[data.Year == i]

    # Build the scatterplot
    plt.scatter(
        x=subsetData['Cum_Publications'],
        y=subsetData['Cum_Citations'],
        s=subsetData['Cum_Citations'] * 15,
        edgecolors="white", linewidth=2, color='midnightblue')

    # Add titles (main and on axis)
    plt.yscale('linear')
    plt.xlabel("Publication")
    plt.ylabel("Citation")
    plt.title("Azad Rasul's Cumulative Publications and Citations during: " + str(i))
    plt.ylim(-50, 500)
    plt.xlim(0, 25)

    # Save the current plot as an image
    buf = io.BytesIO()
    plt.savefig(buf, format='png')
    buf.seek(0)
    images.append(Image.open(buf))

    # Close the figure to avoid memory issues
    plt.close(fig)

# Save the list of images as a GIF
images[0].save('myPubCumulative_animation.gif', save_all=True, append_images=images[1:], optimize=False, duration=200, loop=0)

Webp.net-gifmaker (4).gif

Animated scatterplot

import pandas as pd
import matplotlib.pyplot as plt
from PIL import Image
import io  # Import io module for working with in-memory files

# Read the data (on the web)
data = pd.read_csv('https://raw.githubusercontent.com/holtzy/The-Python-Graph-Gallery/master/static/data/gapminderData.csv')

# Check the first 2 rows
data.head(10)

# And I need to transform my categorical column (continent) into a numerical value group1->1, group2->2...
data['continent'] = pd.Categorical(data['continent'])

# Set the figure size
plt.figure(figsize=(10, 10))

# Subset of the data for year 1952
data1952 = data[data.year == 1952]

# Initialize a list to store images
images = []

# For each year:
for i in data.year.unique():
    # Find the subset of the dataset for the current year
    subsetData = data[data.year == i]

    # Build the scatterplot
    plt.scatter(
        x=subsetData['lifeExp'], 
        y=subsetData['gdpPercap'], 
        s=subsetData['pop']/200000 , 
        c=subsetData['continent'].cat.codes, 
        cmap="Accent", alpha=0.6, edgecolors="white", linewidth=2)

    # Add titles (main and on axis)
    plt.yscale('log')
    plt.xlabel("Life Expectancy")
    plt.ylabel("GDP per Capita")
    plt.title("Year: " + str(i))
    plt.xlim(30, 90)

    # Save the current plot as an image
    buf = io.BytesIO()
    plt.savefig(buf, format='png')
    buf.seek(0)
    images.append(Image.open(buf))

    # Clear the current plot for the next iteration
    plt.clf()

# Save the list of images as a GIF
images[0].save('Gapminder_animation.gif', save_all=True, append_images=images[1:], optimize=False, duration=200, loop=0)

Webp.net-gifmaker (5).gif

Categorical data (catplot)

If the variables are "categorical" (divided into discrete groups), it may be advantageous to use catplot. We can change the plot type by changing the "kind" parameter to violin, swarm, boxen, strip, box, point, bar, or count.

Violin Catplot

Load the titanic.csv file using the load_dataset function from the Seaborn library.

titanic = sns.load_dataset("titanic") # load titanic csv file from seaborn lab

g = sns.catplot(x='pclass', y="age",
                hue="alive",  # catigorize and change the color by alive column data
                data=titanic, kind='violin', legend_out=False) # legend_out = Faluse to move legend to inside the plot
plt.legend(title="Alive", loc=1) # Location: 'upper right':1

image.png

Swarm Catplot

titanic = sns.load_dataset("titanic") # load data

g = sns.catplot(x='pclass', y="age",
                hue="alive",
                data=titanic, kind='swarm', legend_out=False)
plt.axis([-1, 3, 0, 90]) # xlim , ylim
plt.legend(title="Alive", loc=9) # Location: 'upper center':9

image.png

Boxen Catplot

titanic = sns.load_dataset("titanic")
g = sns.catplot(x='pclass', y="age",
                hue="alive",
                data=titanic, kind='boxen', legend_out = False)
plt.axis([-1, 3, 0, 90]) # xlim , ylim
plt.legend(title='Alive', loc = 9)

image.png

Strip Catplot

titanic = sns.load_dataset("titanic")
g = sns.catplot(x='pclass', y="age",
                hue="alive",
                data=titanic, kind='strip', legend_out=False)

plt.axis([-1, 3, 0, 90]) # xlim , ylim
plt.legend(title='Alive', loc = 9)

image.png

Box Catplot

titanic = sns.load_dataset("titanic")
g = sns.catplot(x='pclass', y="age",
                hue="alive",
                data=titanic, kind='box')

image.png

Point Catplot

titanic = sns.load_dataset("titanic")
g = sns.catplot(x='pclass', y="age",
                hue="alive",
                data=titanic, kind='point', legend_out = False)

image.png

Bar Catplot

titanic = sns.load_dataset("titanic")
g = sns.catplot(x='pclass', y="age",
                hue="alive",
                data=titanic, kind='bar', legend_out = False)

image.png

Count Catplot

sns.catplot( x='embark_town',
             kind="count",
             hue= "sex",
             data=titanic,
             height=5,
             aspect=1.5, legend_out = False)
plt.xlabel("Embark town", size=14)
plt.ylabel("Count", size=14)
#plt.tight_layout()
plt.legend(title="Gender", loc=9) # Location: 'upper center':9

image.png

If you find this content helpful, please consider SUBSCRIBING to my channel for future updates.

If you would like to get the full video tutorial and a certificate, you can enroll in the course by following this link: https://www.udemy.com/course/python-for-researchers/?referralCode=886CCF5C552567F1C4E7

Did you find this article valuable?

Support Azad Rasul by becoming a sponsor. Any amount is appreciated!