Python for Data Science (PDS) (3150713)

Unit-04
Data Visualization
 Outline
Looping

✓ Introduction to MatPlotLib
✓ Graph
✓ Plot
✓ Drawing Multiple Lines and Plots
✓ Export graphs/plots to Image/PDF/SVG
✓ Axis, Ticks ad Grids
✓ Line Appearance
✓ Labels, Annotation, Legends
✓ Types of Graphs
✓ Pie Chart
✓ Bar Chart
✓ Histograms
✓ Boxplots
✓ Scatterplots
✓ Time Series
✓ Plotting Geographical data
Introduction to MatPlotLib
 Most people visualize information better when they see it in graphic versus textual format.
 Graphics help people see relationships and make comparisons with greater ease.
 Fortunately, python makes the task of converting textual data into graphics relatively easy
using libraries, one of most commonly used library for this is MatPlotLib.
 Matplotlib is a comprehensive library for creating static, animated, and interactive
visualizations in Python.

#3150713 (PDS)  Unit 04 – Data Visualization 3

Graph
 A Graph or chart is simply a visual representation of numeric data.
 MatPlotLib makes a large number of graph and chart types.
 We can choose any of the common graph such as line charts, histogram, scatter plots etc....

Line Chart Histogram Scatter Plot 3D Plot Images Bar Chart Pie Chart

Etc.......

#3150713 (PDS)  Unit 04 – Data Visualization 4

Plot
 To define a plot, we need some values, the [Link] module and an idea of
what we want to display.
[Link]
1 import [Link] as plt
2 %matplotlib inline
3 values = [5,8,9,4,1,6,7,2,3,8]
4 [Link](range(1,11),values)
5 [Link]()

 In this case, the code tells the [Link]() function to create a plot using x-axis between 1
and 11 and y-axis as per values list.

#3150713 (PDS)  Unit 04 – Data Visualization 5

Plot – Drawing multiple lines
 We can draw multiple lines in a plot by making multiple [Link]() calls.
[Link]
1 import [Link] as plt
2 %matplotlib inline
3 values1 = [5,8,9,4,1,6,7,2,3,8]
4 values2 = [8,3,2,7,6,1,4,9,8,5]
5 [Link](range(1,11),values1)
6 [Link](range(1,11),values2)
7 [Link]()

#3150713 (PDS)  Unit 04 – Data Visualization 6

Plot – Export graphs/plots
 We can export/save our plots on a drive using savefig() method.
[Link]
1 import [Link] as plt
2 %matplotlib inline
3 values1 = [5,8,9,4,1,6,7,2,3,8]
4 values2 = [8,3,2,7,6,1,4,9,8,5]
5 [Link](range(1,11),values1)
6 [Link](range(1,11),values2)
7 #[Link]()
8 [Link]('[Link]',format='png')
[Link]
 Possible values for the format parameters are
 png
 svg
 pdf
 Etc...

#3150713 (PDS)  Unit 04 – Data Visualization 7

Plot – Axis, Ticks and Grid
 We can access and format the axis, ticks and grid on the plot using the axis() method of the
[Link]
[Link]
1 import [Link] as plt
2 %matplotlib notebook
3 values = [5,8,9,4,1,6,7,2,3,8]
4 ax = [Link]()
5 ax.set_xlim([0,50])
6 ax.set_ylim([-10,10])
7 ax.set_xticks([0,5,10,15,20,25,30,35,40,45,50])
8 ax.set_yticks([-10,-8,-6,-4,-2,0,2,4,6,8,10])
9 [Link]()
10 [Link](range(1,11),values)

#3150713 (PDS)  Unit 04 – Data Visualization 8

Plot – Line Appearance
 We need different line styles in order to differentiate when having multiple lines in the same
plot, we can achieve this using many parameters, some of them are listed below.
 Line style (linestyle or ls)
 Line width (linewidth or lw)
 Line color (color or c)
 Markers (marker)
[Link]
1 import [Link] as plt
2 %matplotlib inline
3 values1 = [5,8,9,4,1,6,7,2,3,8]
4 values2 = [8,3,2,7,6,1,4,9,8,5]
5 [Link](range(1,11),values1,c='r',lw=1,ls='--',marker='>')
6 [Link](range(1,11),values2,c='b',lw=2,ls=':',marker='o')
7 [Link]()

#3150713 (PDS)  Unit 04 – Data Visualization 9

Plot – Line Appearance (Cont.)
 Possible Values for each parameters are,
Values Line Style Values Color Values Marker
‘-’ Solid line ‘b’ Blue ‘.’ Point
‘--’ Dashed line ‘g’ Green ‘,’ Pixel
‘-.’ Dash-dot line ‘r’ Red ‘o’ Circle
‘:’ Dotted line ‘c’ Cyan ‘v’ Triangle down
‘m’ Magenta ‘^’ Triangle up
‘y’ Yellow ‘>’ Triangle right
‘k’ Black ‘<’ Triangle left
‘w’ White ‘*’ Star
‘+’ Plus
‘x’ X
Etc.......

#3150713 (PDS)  Unit 04 – Data Visualization 10

Plot – Labels, Annotation and Legends
 To fully document our graph, we have to resort
the labels, annotation and legends.
 Each of this elements has a different purpose
as follows,
 Label : provides identification of a particular data
element or grouping, it will make easy for viewer to Y Label
know the name or kind of data illustrated.
 Annotation : augments the information the viewer
can immediately see about the data with notes,
sources or other useful information.
 Legend : presents a listing of the data groups within
the graph and often provides cues ( such as line
type or color) to identify the line with the data.
Annotation

Legend
X Label

#3150713 (PDS)  Unit 04 – Data Visualization 11

Plot – Labels, Annotation and Legends (Example)
[Link]
1 import [Link] as plt
2 %matplotlib inline
3 values1 = [5,8,9,4,1,6,7,2,3,8]
4 values2 = [8,3,2,7,6,1,4,9,8,5]
5 [Link](range(1,11),values1)
6 [Link](range(1,11),values2)
7 [Link]('Roll No')
8 [Link]('CPI')
9 [Link](xy=[5,1],s='Lowest CPI')
10 [Link](['CX','CY'],loc=4)
11 [Link]()

#3150713 (PDS)  Unit 04 – Data Visualization 12

Choosing the Right Graph
 The kind of graph we choose determines how people view the associated data, so choosing the
right graph from the outset is important.
 For example,
 if we want o show how various data elements contribute towards a whole, we should use pie chart.
 If we want to compare data elements, we should use bar chart.
 If we want to show distribution of elements, we should use histograms.
 If we want to depict groups in elements, we should use boxplots.
 If we want to find patterns in data, we should use scatterplots.
 If we want to display trends over time, we should use line chart.
 If we want to display geographical data, we should use basemap.
 If we want to display network, we should use networkx.
 All the above graphs are there in our syllabus and we are going to cover all the graphs in this
Unit.
 We are also going to cover some other types of libraries which is not in the syllabus like
seaborn, plotly, cufflinks and choropleth maps etc..
#3150713 (PDS)  Unit 04 – Data Visualization 13
Pie Chart
 Pie chart focus on showing parts of a whole, the entire pie would be 100 percentage, the
question is how much of that percentage each value occupies.
[Link]
1 import [Link] as plt
2 %matplotlib notebook
3 values = [305,201,805,35,436]
4 l =
['Food','Travel','Accomodation','Misc','Shoping']
5 c = ['b','g','r','c','m']
6 e = [0,0.2,0,0,0]
7 [Link](values,colors=c,labels=l,explode=e)
8 [Link]()

#3150713 (PDS)  Unit 04 – Data Visualization 14

Pie Chart (Cont.)
 There are lots of other options available with the pie chart, we are going to cover two important
parameters in this slide.
[Link]
1 import [Link] as plt
2 %matplotlib notebook
3 values = [305,201,805,35,436]
4 l =
['Food','Travel','Accomodation','Misc','Shoping']
5 c = ['b','g','r','c','m']
6 [Link](values,colors=c,labels=l,shadow=True,
7 autopct='%1.1f%%')
8 [Link]()

#3150713 (PDS)  Unit 04 – Data Visualization 15

Bar charts
 Bar charts make comparing values easy, wide bars an d segregated measurements emphasize
the difference between values, rather that the flow of one value to another as a line graph.
[Link]
1 import [Link] as plt
2 %matplotlib notebook
3 x = [1,2,3,4,5]
4 y = [5.9,6.2,3.2,8.9,9.7]
5 l = ['1st','2nd','3rd','4th','5th']
6 c = ['b','g','r','c','m']
7 w = [0.5,0.6,0.3,0.8,0.9]
8 [Link]('Sem wise spi')
9 [Link](x,y,color=c,label=l,width=w)
10 [Link]()

#3150713 (PDS)  Unit 04 – Data Visualization 16

Histograms
 Histograms categorize data by breaking it into bins, where each bin contains a subset of the
data range.
 A Histogram then displays the number of items in each bin so that you can see the distribution
of data and the progression of data from bin to bin.
[Link]
1 import [Link] as plt
2 import numpy as np
3 %matplotlib notebook
4 cpis = [Link](0,10,100)
5 [Link](cpis,bins=10,
histtype='stepfilled',align='mid',label
='CPI Hist')
6 [Link]()
7 [Link]()

#3150713 (PDS)  Unit 04 – Data Visualization 17

Boxplots
 Boxplots provide a means of depicting groups of numbers through their quartiles.
 Quartiles means three points dividing a group into four equal parts.
 In boxplot, data will be divided in 4 part using the 3 points (25th percentile, median, 75th
percentile)
Interquartile Range
(IQR)

Outliers Whiskers Whiskers Outliers

Minimum Maximum
(Q1 – 1.5 * IQR) Median (Q3 + 1.5 * IQR)

Q1 Q2 Q3
(25th Percentile) (50th Percentile) (75th Percentile)

-5 -4 -3 -2 -1 0 1 2 3 4 5
#3150713 (PDS)  Unit 04 –Data Visualization 18
Boxplot (Cont.)
 Boxplot basically used to detect outliers in the data, lets see an example where we need
boxplot.
 We have a dataset where we have time taken to check the paper, and we want to find the
faculty which either takes more time or very little time to check the paper.
[Link]
1 import pandas as pd
2 import [Link] as plt
3 %matplotlib inline
4 timetaken =
[Link]([50,45,52,63,70,21,56,68,54,5
7,35,62,65,92,32])
5 [Link](timetaken)

 We can specify other parameters like

 widths, which specify the width of the box
 notch, default is False
 vert, set to 0 if you want to have horizontal graph
#3150713 (PDS)  Unit 04 – Data Visualization 19
Scatter Plot
 A scatter plot is a type of plot that shows the data as a collection of points.
 The position of a point depends on its two-dimensional value, where each value is a position on
either the horizontal or vertical dimension.
 It is really useful to study the relationship/pattern between variables.
[Link]
1 import [Link] as plt
2 import pandas as pd
3 %matplotlib inline
4 df = pd.read_csv('[Link]')
5 [Link](df['bmi'], df['charges'])
6 [Link]()

#3150713 (PDS)  Unit 04 – Data Visualization 20

Scatter Plot (Cont.)
 To find specific pattern from the data, we can further divide the data and plot scatter plot.
 We can do this with the help of groupby method of DataFrame, and then using tuple unpacking
while looping the group.
[Link]
1 import [Link] as plt
2 import pandas as pd
3 %matplotlib inline
4 df = pd.read_csv('[Link]')
5 grouped = [Link](['smoker'])
6 for key, group in grouped:
7 [Link](group['bmi'],
group['charges'],
label='Smoke = '+key)
8 [Link]()
9 [Link]()
 Note : we can specify marker, color, and size of the marker with the help
of marker, color and s parameter respectively.
#3150713 (PDS)  Unit 04 – Data Visualization 21
Time Series
 Observations over time can be considered as a Time Series.
 Visualization plays an important role in time series analysis and forecasting.
 Time Series plots can provide valuable diagnostics to identify temporal structures like trends,
cycles, and seasonality.
 In order to create a Time Series we first need to get the date range, which can be created with
the help of datetime and pandas library.
[Link] OUTPUT
1 import pandas as pd DatetimeIndex(['2020-08-28', '2020-
08-29', '2020-08-30', '2020-08-31',
2 import datetime as dt
'2020-09-01', '2020-09-02', '2020-
3 start_date = [Link](2020,8,28) 09-03', '2020-09-04', '2020-09-
4 end_date = [Link](2020,9,05) 05'],
5 daterange = pd.date_range(start_date,end_date) dtype='datetime64[ns]', freq='D')
6 print(daterange)

#3150713 (PDS)  Unit 04 – Data Visualization 22

Time Series (Cont.)
 We can use some more parameters for date_range() function like
 freq, to specify the frequency at which we want the date range (default is ‘D’ for days)
 periods, number of periods to generate in between start/end or from start with freq.
 We can also create a date range with the help of startdate, periods and freq, for example
[Link] OUTPUT
1 import pandas as pd DatetimeIndex(['2020-08-25', '2020-
08-26', '2020-08-27', '2020-08-28',
2 import datetime as dt
'2020-08-29', '2020-08-30', '2020-
3 start_date = [Link](2020,8,28) 08-31', '2020-09-01', '2020-09-02',
4 daterange = '2020-09-03'],
5 pd.date_range(start_date,freq='D',periods=10) dtype='datetime64[ns]', freq='D')
6 print(daterange)
 Some of important possible values for the freq are
 D, for calendar day  H, for hour  B, for business day
 W, for week  T/min, for minute  SM, for semi month end
 M, for month  S, for seconds  Q, for quarter end
 Y, for year  L, for milliseconds  BQ, for business quarter end
#3150713 (PDS)  Unit 04 – Data Visualization 23
Basemap
 The matplotlib basemap toolkit is a library for plotting 2D data on maps in Python.

#3150713 (PDS)  Unit 04 – Data Visualization 24

NetworkX
 We can use networkx library in order to deal with any kind of networks, which includes social
network, railway network, road connectivity etc….
 Install
 pip install networkx
 conda install networkx
 Types of network graph
 Undirected
 Directed
 Weighted graph

#3150713 (PDS)  Unit 04 – Data Visualization 25

NetworkX (example)
[Link]
1 import networkx as nx
2 g = [Link]() # undirected graph
3 g.add_edge('rajkot','junagadh')
4 g.add_edge('junagadh','porbandar')
5 g.add_edge('rajkot','jamnagar')
6 g.add_edge('jamnagar','bhanvad')
7 g.add_edge('bhanvad','porbandar')
8 [Link](g,with_labels=True)

[Link]
1 import networkx as nx
2 gD = [Link]() # directed graph
3 gD.add_edge('Modi','Arjun')
4 gD.add_edge('Modi','GambhavaSir')
5 gD.add_edge('GambhavaSir','Modi')
6
7 [Link](gD, with_labels=True)
#3150713 (PDS)  Unit 04 –Data Visualization 26
NetworkX (cont.)
 We can use many analysis functions available in NetworkX library, some of functions are as
below
 nx.shortest_path(g,'rajkot','porbandar')
▪ Will return ['rajkot', 'junagadh', 'porbandar']
 [Link](g)
▪ Will return clustering value for each node
 nx.degree_centrality(g)
▪ Will return the degree of centrality for each node, we can find most
popular/influential node using this method.
 [Link](g)
▪ Will return the density of the graph.
▪ The density is 0 for a graph without edges and 1 for a complete graph.
 [Link](g)
▪ Return a summary of information for the graph G.
▪ The summary includes the number of nodes and edges, and their average degree.

#3150713 (PDS)  Unit 04 – Data Visualization 27