18
Harvard IACS CS109A Pavlos Protopapas, Kevin Rader, and Chris Tanner A vehicle for EDA and communication Lecture 14: Visualization

Lecture 14: Visualization - GitHub Pages

  • Upload
    others

  • View
    5

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Lecture 14: Visualization - GitHub Pages

Harvard IACSCS109APavlos Protopapas, Kevin Rader, and Chris Tanner

A vehicle for EDA and communication

Lecture 14: Visualization

Page 2: Lecture 14: Visualization - GitHub Pages

As the matplotlib thickens …

Page 3: Lecture 14: Visualization - GitHub Pages

CS109A, PROTOPAPAS, RADER, TANNER 3

• Homework 3 is due Oct 7 (Wed) @ 11:59pm

• Didn’t submit Milestone 1? Want to change project selection?E-mail the helpline at [email protected]

• After lecture, please update your Zoom to the latest version (5.3.1).

ANNOUNCEMENTS

Page 4: Lecture 14: Visualization - GitHub Pages

• Understand why visualization/plotting is important

• Learn aspects that tend to make visualizes effective and ineffective

• Feel comfortable designing plots that best convey your message

• Gain experience in producing plots with Python

4

Learning Objectives

be more cognizant of broader design choices(e.g., typography, s p a c i n g , colors) Extra Goal

Page 5: Lecture 14: Visualization - GitHub Pages

5

Agenda

CommunicationMotivation

Key Considerations

EDA Refresher

Effective VisualizationGraphical Integrity

Scope

Displays

Sensible Design

Page 6: Lecture 14: Visualization - GitHub Pages

6

Agenda

CommunicationMotivation

Key Considerations

EDA Refresher

Effective VisualizationGraphical Integrity

Scope

Displays

Sensible Design

Page 7: Lecture 14: Visualization - GitHub Pages

Assume you know a given dataset is credible,

complete with the info you want, and has no missing values.

7

Why do further EDA?

Page 8: Lecture 14: Visualization - GitHub Pages

Purposes of EDA:• Maximize insight into a dataset

• Uncover underlying structure

• Detect outliers

• Test underlying assumptions

• Develop parsimonious models

8

Page 9: Lecture 14: Visualization - GitHub Pages

CS109A, PROTOPAPAS, RADER, TANNER 9

EDA Refresher: English Premier League

name club age position market value

Alexis Sanchez

Mesut Ozil

Petr Cech

Theo Walcott

Laurent Koscielny

Arsenal

Arsenal

Arsenal

Arsenal

Arsenal

28

28

35

28

31

LW

AM

GK

RW

CB

65

50

7

20

22

from www.transfermarkt.us

Page 10: Lecture 14: Visualization - GitHub Pages

CS109A, PROTOPAPAS, RADER, TANNER 10

EDA Refresher: English Premier League

Page 11: Lecture 14: Visualization - GitHub Pages

CS109A, PROTOPAPAS, RADER, TANNER 11

EDA Refresher: English Premier League

Page 12: Lecture 14: Visualization - GitHub Pages

CS109A, PROTOPAPAS, RADER, TANNER 12

EDA Refresher: English Premier League

Are the outliers legit?

Page 13: Lecture 14: Visualization - GitHub Pages

CS109A, PROTOPAPAS, RADER, TANNER 13

EDA Refresher: English Premier League

Page 14: Lecture 14: Visualization - GitHub Pages

CS109A, PROTOPAPAS, RADER, TANNER 14

EDA Refresher: English Premier League

Page 15: Lecture 14: Visualization - GitHub Pages

CS109A, PROTOPAPAS, RADER, TANNER 15

EDA Refresher: English Premier League

Page 16: Lecture 14: Visualization - GitHub Pages

16

Agenda

CommunicationMotivation

Key Considerations

EDA Refresher

Effective VisualizationGraphical Integrity

Scope

Displays

Sensible Design

DISCLAIMER. Some of these examples involve political data. In no way should this be taken as a signal of my support or endorsement in any beliefs; the point is merely to convey good and bad choices when it comes to effective visualization.

Page 17: Lecture 14: Visualization - GitHub Pages

17

Agenda

CommunicationMotivation

Key Considerations

EDA Refresher

Effective VisualizationGraphical Integrity

Scope

Displays

Sensible Design

DISCLAIMER. Some of these examples involve political data. In no way should this be taken as a signal of my support or endorsement in any beliefs; the point is merely to convey good and bad choices when it comes to effective visualization.

Page 18: Lecture 14: Visualization - GitHub Pages

18

Agenda

CommunicationMotivation

Key Considerations

EDA Refresher

Effective VisualizationGraphical Integrity

Scope

Displays

Sensible Design

DISCLAIMER. Some of these examples involve political data. In no way should this be taken as a signal of my support or endorsement in any beliefs; the point is merely to convey good and bad choices when it comes to effective visualization.