ENG 202: Computers and Engineering Object Oriented Programming in PYTHON

ENG 202: Computers and Engineering
Object Oriented Programming in PYTHON

LECTURE 12 – Data Visualization in PYTHON
Maurice J. KHABBAZ, Ph.D.

Visualizing Results
• Throughout so many upcoming engineering courses:
• One quantifies system performances using different metrics:
• Examples:
• Electrical Engineering:
• Voltage (current) across (through) a resistive element – Ohm’s Law:
𝑉𝑉 = 𝑅𝑅 × 𝐼𝐼
• 𝑉𝑉 ≜ Voltage ; 𝑅𝑅 ≜ Resistance ; 𝐼𝐼 ≜ Current intensity
• Mechanical Engineering:
• Energy used by a system to do useful work under fixed pressure, temperature and volume:
𝐺𝐺 = 𝑈𝑈 + 𝑝𝑝𝑝𝑝 − 𝑇𝑇𝑇𝑇
• 𝐺𝐺 ≜ Gibbs Free Energy ; 𝑈𝑈 ≜ Internal Energy ; 𝑝𝑝 ≜ Pressure ; 𝑇𝑇 ≜ Temperature ; 𝑆𝑆 ≜ Entropy
• Civil Engineering:
• In Traffic Flow Theory, the average speed of a vehicle navigating along a roadway segment is:
𝜌𝜌
𝑉𝑉 = 𝑉𝑉FF 1 −
𝜌𝜌max
• 𝑉𝑉𝐹𝐹𝐹𝐹 ≜ Free-Flow Speed ; 𝜌𝜌 ≜ Traffic Density ; 𝜌𝜌max ≜ maximum (jamming) traffic density.
Friday, January 3, 2020 NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. 2

Visualizing Results
• Use graphs to provide an intuitive/visual sense of system behavior.
• This is yet another example on leveraging existing libraries
• Rather than writing procedures from scratch.
• PYTHON provides libraries for (among other topics):
• Graphing.
• Numerical Computation.
• Stochastic Computation.
• Explore the usage of existing library procedures for:
• Processing data.
• Exploring data.
Using matplotlib
• Most popular PYTHON library for data visualization.
• Written in PYTHON and makes use of numpy.
• Cross-platform library for making 2D plots from data arrays.
• Provides Object-Oriented Application Programming Interface (OO-API):

• Helps embed plots in applications using Graphical User Interface (GUI) toolkits.
• Can be used in PYTHON and interactive PYTHON (iPython) shells (and others).
• Has a procedural interface called Pylab (designed to resemble MATLAB):

• matplotlib combined with numpy are the open-source equivalent of MATLAB.

Using matplotlib
• matplotlib.pyplot is a collection of command style functions:
• Make matplotlib work like MATLAB.
• Each function:
• Creates a figure.
• Creates a plotting area in the figure.
• Plots some lines in the plotting area.
• Decorates the plot with labels, … etc.

TYPE OF PLOT DESCRIPTION
bar Creates a bar plot
barh Creates a horizontal bar plot
boxplot Creates a box and a whisker plot
hist Plots a histogram
hist2d Plots a 2D histogram
pie Plots a pie chart
Types of Plots plot Plots lines and/or markers on axes
polar Creates a polar plot
scatter Creates a scatter plot of x VS y
stackplot Creates a stacked area plot
stem Creates a stem plot
step Creates a step plot

Friday, January 3, 2020
quiver Plots a 2D field of arrows
NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. 6
Image Functions
Function Description
imread Read an image from a file into an array
imsave Save an array as an image file
imshow Display an image on the axes

Axis Functions
FUNCTION DESCRIPTION
axes Add axes to the figure
text Add text to the axes
title Set a title for the current axes
xlabel Set the x-axis label for the current axes
xlim Get or set the x limits of the current axes
xscale Set the scaling of the x-axis for the current axes
xticks Get or set the x limits of the current tick locations and labels
ylabel Set the y-axis label for the current axes
ylim Get or set the y limits for the current axes
yscale Set the scaling of the y-axis for the current axes
yticks Get or set the y limits of the current tick locations and labels

Figure Functions
Function Description
figtext Add text to figure
figure Creates a new figure
show Displays a figure
savefig Saves the current figure
close Closes a figure window

A Simple Plot
• First, need to import library into the computing environment:
import matplotlib as mpl
• Allows referencing any of the library’s functions as: mpl.<funcName>

• Provides access to existing set of graphing/plotting functions.
• This is just a simple example.
• Lots of additional information is available in matplotlib documentation.
• Will go over many other examples in details later in this lecture.
A Simple Plot – Cont’d.
• Basic function to plot lists as x and y values:
• Other data structures are much more powerful.
• Example:
from matplotlib import pyplot as plt
x = []; y = [];
for v in range(31):
x.append(v);
y.append(v);
plt.plot(x, y) Sample Display
plt.show()

A Simple Example: Adding More Stuff
• Example:
import numpy as np
x = []; yLinear = []; yQuad = [];
yCubic = []; yExp = []
for v in range(10):
x.append(v);
yLinear.append(v); yQuad.append(v**2)
yCubic.append(v**3); yExp.append((np.e)**v)
plt.plot(x, yLinear); plt.plot(x, yQuad)
plt.plot(x, yCubic); plt.plot(x, yExp)
plt.show()

A Simple Example: Adding More Stuff
Sample Display • Four graphs overlaid.

• Guess red curve is the exponential:
• Grows really much faster than the others.
• Green curve is most likely the cubic one.
• Brown curve is the quadratic one.
• Blue curve is the linear one.
• Brown and Blue curves merely visible:
• This is because of the scale of the y-axis:
for v in range(10):
• y-axis grows to 8000.

Overlapping Displays
• Previous display unhelpful since only the biggest plot is clearly visible:
• Due to the difference in the scales of the different curves.
• Is it possible to graph each function separately?
• Yes, the function plt.figure() makes this possible.
• Syntax:
plt.figure(<arg>)
• Creates a new display with that name if one does not already exist.
• If a display with that name exists, it is reopened for further processing.

Separating Plots
import numpy as np
x = []; yLinear = []; yQuad = [];
yCubic = []; yExp = []
for v in range(10):
x.append(v);
yLinear.append(v); yQuad.append(v**2)
yCubic.append(v**3); yExp.append((np.e)**v)
plt.figure(“Linear”); plt.plot(x, yLinear)
plt.figure(“Quadratic”); plt.plot(x, yQuad)
plt.figure(“Cubic”); plt.plot(x, yCubic);
plt.figure(“Exponential”); plt.plot(x, yExp)
plt.show()
Separating Plots: Sample
Displays

Plotting: Adding More and More Stuff
• A professional plot must have well labelled axes:
• Can be done using the plt.xlabel() and plt.ylabel() functions.
• Example:
import numpy as np
x = []; yLinear =
for v in range(10):
x.append(v);
yLinear.append(v);
plt.xlabel(“x Values”); plt.ylabel(“Linear Function”)
plt.show()
Adding
Labels to the
Axes

Plotting: Adding a Title to The Graph
• A professional plot must have well labelled axes:
• Can be done using the plt.xlabel() and plt.ylabel() functions.
• Example:
import numpy as np
x = []; yLinear =
for v in range(10):
x.append(v);
yLinear.append(v);
plt.xlabel(“x Values”); plt.ylabel(“Linear Function”)
plt.title(“Plot of a Linear Function”)
plt.show()
Adding a
Title to The
Graph

Cleaning a Figure Window
• This means erasing a figure window with all its custom set properties:
• i.e. axes labels, titles, curve colors, etc.
• This can be done using the plt.clf() function.
• Example:
plt.clf()

Comparing Results
• Suppose it is required to compare different plots.
• The scales on the graphs are very different.
• Option 1: explicitly set limits on one or both axes.
• Option 2: plot multiple function on the same display.

Option 1: Changing Limits of One or Both Axes
plt.figure(“Linear”); plt.clf()
plt.ylim(0, 1000)
plot(x, yLinear)
plt.title(“Plot of Linear Function”)
plt.figure(“Quadratic”); plt.clf()
plt.ylim(0, 1000)
plt.plot(x, yQuad)
plt.title(“Plot of Quadratic Function”)

Option 1: Sample Display

Option 2: Overlaying Plots
plt.figure(“LinQuad”); plt.clf()
plt.plot(x, yLinear); plt.plot(x, yQuad)
plt.title(“Linear V.S. Quadratic Functions”)
plt.figure(“CubExp”); plt.clf()
plt.plot(x, yCubic); plt.plot(x, yExp)
plt.plot(“Cubic V.S. Exponential Functions”)

Option 2: Sample Display
• Important message here:

• In comparing, one needs to think about what is
being compared to what.
• Python gives this nice way of visualization.
• Sometimes:
• Like to overlay things on the same window.
• Like to display things side by side.

Plotting: Adding More Documentation
• A legend helps identifying which of overlaying plots is which.
• Example:
plt.plot(x, yLinear, label = “Linear Function”)
plt.plot(x, yQuad, label = “Quadratic Function”)
plt.legend(loc = “upper right”)specify legend location in figure window
plt.plot(x, yCubic, label = “Cubic Function”)
plt.plot(x, yExp, label = “Exponential Function”)
plt.legend() not specifying location lets Python decide the best location
Sample Display With Legend

Plotting: Controlling Display Parameters
Requirement: control details of the displays themselves.
Examples:
Change color or style Change width Subdividing a figure window
of data sets. of lines or displays. into multiple subplots.
Controlling Display Parameters – Cont’d.
plt.plot(x, yLinear, “b-”, label = “Linear Function”)
plt.plot(x, yQuad, “ro”, label = “Quadratic Function”)
plt.legend(loc = “upper right”)
plt.plot(x, yCubic, “g^”, label = “Cubic Function”)
plt.plot(x, yExp, “r--”, label = “Exponential Function”)
plt.legend()

Changing Data Sample Display

Changing Line Width
plt.plot(x, yLinear, “b-”, label = “Linear Function”, linewidth = 2.0)
plt.plot(x, yQuad, “ro”, label = “Quadratic Function”, linewidth = 3.0)
plt.legend(loc = “upper right”)
plt.plot(x, yCubic, “g^”, label = “Cubic Function”, linewidth = 4.0)
plt.plot(x, yExp, “r--”, label = “Exponential Function”, linewidth = 5.0)
plt.legend()

Changing Line Width Sample Display

Plotting: Using Subplots
• Subplotting ⇒ subdividing a figure window into a matrix of figures.
• Syntax:
plt.subplot(xyz)
• where x, y and z are integers designating the following:

• x ≜ number of rows.
• y ≜ number of columns.
• z ≜ number of the figure to be handled.

Example: Using Subplots
plt.figure("LinQuad")
plt.clf()
plt.subplot(211); plt.ylim(0, 80); plt.ylabel("Linear Function")
plt.plot(x, yLinear, "b-", label = "Linear Function", linewidth = 2.0);
plt.subplot(212); plt.ylim(0,80)
plt.ylabel("Quadratic Function"); plt.xlabel("x Values")
plt.plot(x, yQuad, "r", label = "Quadratic Function", linewidth = 3.0)
plt.figure("CubExp")
plt.clf()
plt.subplot(121); plt.ylim(0, 8000);
plt.ylabel("Cubic Function"); plt.xlabel("x Values")
plt.plot(x, yCubic, "g--", label = "Cubic Function", linewidth = 4.0)
plt.subplot(122); plt.ylim(0, 8000)
plt.ylabel("Exponential Function"); plt.xlabel("x Values")
plt.plot(x, yExp, "r", label = "Exponential Function", linewidth = 5.0)
Sample Display Using Subplots

Plotting: Changing Scales
plt.figure(“CubExp-Y-LogScale")
plt.clf()
plt.yscale(“log”)
plt.legend(); plt.title(“Cubic V.S. Exponential Functions (y-Log)”)
plt.xlabel(“x Values”)
plt.figure("CubExp-Y-Linear")
plt.clf()
plt.legend(); plt.title(“Cubic V.S. Exponential Function (y-Linear)”)
plt.xlabel(“x Values”)

Sample Display with Changing Scales

Additional tutorial-oriented documentation is found at:
Very
Useful
Tutorial
https://www.tutorialspoint.com/matplotlib/index.htm

Practical Example: Analysis V.S. Contribution
• Objectives:
• Illustrate how plotting allows the proper data visualization.
• Show how the visualization of results helps in:
• Guiding computations.
• Determining the values of parameters to pass as inputs to perform computations.
• Enumerate things to explore within/throughout certain computations.
• Problem Statement:
• Imagine planning a retirement.
• Think about ensuring to have enough savings to allow retiring at a certain age:
• Intend to save an amount m each month.
• Expect to earn a percentage r of income on investments each month (i.e. interest)
• Have a look at how a retirement fund grows with compounding of interest:
• Given m and r, determine the time when ready for retirement.

Practical Example: Computing Compound Interest
• Write a function called retire() that takes in three arguments:
• monthly ≜ the amount to put aside each month.
• rate ≜ the rate a which income grows as a result of investments.
• terms ≜ the period of time to do the computation over.
• The function should set up two lists, namely: base and savings
• These are going to be the x and y values.
• The monthly rate is equivalent to the yearly rate divided by twelve.
• Savings at current month are equivalent to:
• The latest accumulated savings to which is accumulated the monthly rate
• Plus the monthly amount to be put aside.
• Finally return the base and savings together in a tuple.

Practical Example: Computing Compound Interest
def retire(monthly, rate, terms):
savings = [0]
base = [0]
monthlyRate = rate / 12.0
for i in range(terms):
base += [i]
savings += [savings[-1] * (1 + monthlyRate) + monthly]
return (base, savings)

Practical Example: Display Results V.S. Months
• Write a function displayResults() that takes three arguments:
• monthlies ≜ a list of trial monthly amounts to put aside in [500; 1100].
• rate ≜ the yearly rate of income as before.
• terms ≜ the number of months to look over, also, as before.
• The function will then plot the different curves that correspond to
each and every trial monthly amount put aside.
• This will allow:
• The comparison of results.
• Determination of best plan to follow given: i) interest and ii) retirement time.

Practical Example: Display Results V.S. Months
def displayResults(monthlies, rate, terms):
plt.figure(“retirementChart”)
plt.clf()
for monthly in monthlies:
(x, y) = retire(monthly, rate, terms)
plt.plot(x, y, label = “Monthly Amount: ” + str(monthly))
plt.legend(loc = “upper left”)
plt.xlabel(“Number of Months”); plt.ylabel(“Retirement Amount”);
plt.show()
# Main Program: # Do not forget from matplotlib import pyplot as plt
displayResults([500, 600, 700, 800, 900, 1000, 1100], 0.05, 40 * 12)

Practical
Example:
Sample
Display
• Can see the impact of increasing monthly contribution
45 • Retirement amount ranges from $750K to $1.67M as monthly savings

ranges from $500 to $1100.
• Question: What is the effect of rate of growth of investments?
NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. Friday, January 3, 2020

Practical Example: Display Results V.S. Rate
• Modify displayResults() to take in a:
• Fixed monthly saving amount.
• List of different income rates.
• Do the plotting to visualize the variations of the retirement amount as

a result of varying the income rates.

Practical Example: Display Results V.S. Rate
def displayResults(monthly, rates, terms):
plt.clf()
for rate in rates:
plt.plot(x, y, label = “Monthly Amount: ” + str(monthly) +
“, Rate: ” + str(int(rate * 100)) + “%”)
plt.show()
displayResults(800, [0.03, 0.05, 0.07], 40 * 12)

Practical
Example:
Sample
Display
• Varying growth of investment return rate spreads the graphs more rapidly.
48
• Aggressive growth (optimistic 7%) leads to better amounts upon retirement.
• Retirement amount ∈ [$600K; $2.1M] as rate grows in [3%; 7%].
• Question: What if change affects both monthly savings and rate together?

Practical Example: Display Results V.S. Both
def displayResults(monthlies, rates, terms):
plt.clf(); plt.xlim(30 * 12, 40 * 12)
for monthly in monthlies:
for rate in rates:
plt.plot(x, y, label = “Monthly Amount: ” + str(monthly) +
“, Rate: ” + str(int(rate * 100)) + “%”)
plt.show()
displayResults([500, 700, 900], [0.03, 0.05, 0.07], 40 * 12)

Practical
Example:
Sample
Display
• Labels are kind of helpful in telling what is going on.

• Zooming in on just the last ten years is valuable as it spreads out curves.
50 •
•
A bit hard to figure out which of these graphs corresponds to which color.
Need to fix this to really see more appropriately the comparison.
• Objective: How to use visualization to help decide how to think about computation?

Practical
Example:
Sample
Display
• Too much overlay of graphs that does not allow distinguishing among them.
51
• Question: How to separate those out into pieces that help think better?
• Maybe analyze them separately: be careful about visualizing different effects.
• Do that as follows on the next slide.

Practical Example: Display Results V.S. Both
def displayResults(monthlies, rates, terms):
plt.clf(); plt.xlim(30 * 12, 40 * 12)
monthLabels = [“r”, “b”, “g”, “k”]; rateLabels = [“-”, “o”, “--”]
for i in range(len(monthlies)):
monthly = monthlies[i]; monthLabel = monthLabels[i % len(monthLabels)]
for j in range(len(rates)):
rate = rates[j]; rateLabel = rateLabels[j % len(rateLabels)]
plt.plot(x, y, monthLabel + rateLabel, label = “Monthly Amount: ” + str(monthly) +
“, Rate: ” + str(int(rate * 100)) + “%”)
plt.show()
displayResults([500, 700, 900], [0.03, 0.05, 0.07], 40 * 12)

Practical
Example:
Sample
Display
• Notice there are now only three different colors and styles.
• Red is all $500, Blue is all $700 and Green is all $900.
53 • The color indicates very nicely the monthly savings amount.

• The style (i.e. solid line, circle, dashed line) indicates the return rate.
• There is still a bit overlay but now graphs are much easier to read.
Practical
Example:
Sample
Display
• The four dashed lines are much better than the slant of the four solid lines.
• Objective realized: visualize and use visualization to guide computation.
54 • Start saving early for retirement to get wonderful compound interest.

• Much better than relying on getting 25% return each year.
• Do not invest in a Ponzi scheme as this is not a good idea.
Summary
• Now easier to see grouping plots:
• Color encodes monthly contributions.
• Format/style encodes growth rate of investments.
• Interaction with plotting routines and computations allows data exploration:
• Change display range to zoom in on particular areas of interest.
• Change value sets and visualize effect – then guide new choice of values to explore.
• Change display parameters to highlight clustering of plots by parameter.
• There are still so many other examples about using plotting to visualize data.
• Example of using libraries:
• All what has been done throughout this lecture is usage of a set of functions.
• These functions are already present in matplotlib.
• That makes it possible to rely on somebody having written code that can be reused.
• Libraries are tools that will be used a lot as one moves on with own computation.

ENG 202: Computers and Engineering Object Oriented Programming in PYTHON

Uploaded by

ENG 202: Computers and Engineering Object Oriented Programming in PYTHON

Uploaded by

ENG 202: Computers and Engineering

Object Oriented Programming in PYTHON

Maurice J. KHABBAZ, Ph.D.

Friday, January 3, 2020 NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. 2

• Written in PYTHON and makes use of numpy.

• Cross-platform library for making 2D plots from data arrays.

• Provides Object-Oriented Application Programming Interface (OO-API):

• Has a procedural interface called Pylab (designed to resemble MATLAB):

Friday, January 3, 2020 NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. 4

Friday, January 3, 2020 NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. 5

bar Creates a bar plot

barh Creates a horizontal bar plot

boxplot Creates a box and a whisker plot

hist Plots a histogram

hist2d Plots a 2D histogram

pie Plots a pie chart

Types of Plots plot Plots lines and/or markers on axes

polar Creates a polar plot

scatter Creates a scatter plot of x VS y

stackplot Creates a stacked area plot

stem Creates a stem plot

step Creates a step plot

imread Read an image from a file into an array

imsave Save an array as an image file

imshow Display an image on the axes

Friday, January 3, 2020 NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. 7

Friday, January 3, 2020 NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. 8

figtext Add text to figure

figure Creates a new figure

show Displays a figure

savefig Saves the current figure

close Closes a figure window

Friday, January 3, 2020 NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. 9

import matplotlib as mpl

• Allows referencing any of the library’s functions as: mpl.<funcName>

Friday, January 3, 2020 NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. 11

Friday, January 3, 2020 NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. 12

Sample Display • Four graphs overlaid.

Friday, January 3, 2020 NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. 13

Friday, January 3, 2020 NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. 14

Friday, January 3, 2020 NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. 16

Friday, January 3, 2020 NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. 18

Friday, January 3, 2020 NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. 20

Friday, January 3, 2020 NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. 21

• The scales on the graphs are very different.

• Option 1: explicitly set limits on one or both axes.

• Option 2: plot multiple function on the same display.

Friday, January 3, 2020 NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. 22

Friday, January 3, 2020 NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. 23

Friday, January 3, 2020 NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. 24

Friday, January 3, 2020 NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. 25

• Important message here:

Friday, January 3, 2020 NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. 26

Friday, January 3, 2020 NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. 28

Requirement: control details of the displays themselves.

Friday, January 3, 2020 NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. 30

Friday, January 3, 2020 NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. 31

Friday, January 3, 2020 NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. 32

Friday, January 3, 2020 NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. 33

• where x, y and z are integers designating the following:

Friday, January 3, 2020 NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. 34

Friday, January 3, 2020 NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. 36

Friday, January 3, 2020 NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. 37

Friday, January 3, 2020 NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. 38

Friday, January 3, 2020 NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. 39

Friday, January 3, 2020 NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. 40

Friday, January 3, 2020 NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. 41

Friday, January 3, 2020 NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. 42

Friday, January 3, 2020 NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. 43

Friday, January 3, 2020 NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. 44

• Can see the impact of increasing monthly contribution

45 • Retirement amount ranges from $750K to $1.67M as monthly savings

NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. Friday, January 3, 2020

• Do the plotting to visualize the variations of the retirement amount as

Friday, January 3, 2020 NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. 46

Friday, January 3, 2020 NDU - ENG 202 - Maurice J. KHABBAZ, Ph.D. 47