One Fourth Labs
We deliver courseware in AI and related areas
Data Visualisation
collect
process
store
describe
model
Recap: What is Data Science?
Describing data is an act of compressing the information to focus on
One way is to compute statistics
The other way is to visualise the data
Data visualisation - What
visual encoding of data
Data visualisation - What
visual encoding of data
This encoding can be identified by us,
but can be decoded only by machines
What we are interested in are representations that are decoded by us
Data visualisation - What
visual elements for encoding data
We make perceptual errors in decoding these elements
Data visualisation - What
Data visualisation - What
A
B
Our goal is to reduce perceptual error
Which is longer A or B?
Data visualisation - What
A
B
Our goal is to reduce perceptual error
Which is longer A or B?
Data visualisation - Why
discover insights from the data
RangeIndex: 53940 entries, 0 to 53939
Data columns (total 10 columns):
# Column Non-Null Count Dtype
--- ------ -------------- -----
0 carat 53940 non-null float64
1 cut 53940 non-null object
2 color 53940 non-null object
3 clarity 53940 non-null object
4 depth 53940 non-null float64
5 table 53940 non-null float64
6 price 53940 non-null int64
7 x 53940 non-null float64
8 y 53940 non-null float64
9 z 53940 non-null float64
dtypes: float64(6), int64(1), object(3)
memory usage: 4.1+ MB diamonds dataset
Data visualisation - Why
discover insights from the data
communicate insights effectively
Data visualisation - Why
discover insights from the data
communicate insights effectively
Survey of how Americans spend their day
by NYTimes
Animated gapminder plot by Hans Rosling
Pitfalls in Data Visualisation
It is not that important
Pitfalls in Data Visualisation
It is about packing information in an image
Pitfalls in Data Visualisation
It is about drawing cool images
Pitfalls in Data Visualisation
It is about drawing cool images
Pitfalls in Data Visualisation
Aesthetic features like colour don't matter
Pitfalls in Data Visualisation
Limited number of courses on
data visualisation
Important to figure out
what and how to learn
What to learn
Syntax of plotting libraries
Understanding of plot types and usages
Aesthetics of plots
Communicating with plots
What to learn
Syntax of plotting libraries
Understanding of plot types and usages
Aesthetics of plots
Communicating with plots
Syntax is not hard to learn, or Google
Many small details can be easily searched
What to learn
Syntax of plotting libraries
Understanding of plot types and usages
Aesthetics of plots
Communicating with plots
Most crucial to develop a structured approach to thinking about plots
When to use which plot type, why?
What to learn
Syntax of plotting libraries
Understanding of plot types and usages
Aesthetics of plots
Communicating with plots
Clever people have been thinking about this
Good libraries have aesthetic default choices
What to learn
Syntax of plotting libraries
Understanding of plot types and usages
Aesthetics of plots
Communicating with plots
Very important to use plots as an element in communication (research paper, presentation, or blog)
Our Focus
Syntax of plotting libraries
Understanding of plot types and usages
Aesthetics of plots
Communicating with plots
We will focus on understanding plot types from first principles and then seeing examples of their usage
Pay Attention To
How to tradeoff readability & information
What are the different plot elements & how do they compose
How to re-orient thinking from data-centric to viewer-centric objectives
Package
MATPLOTLIB
SEABORN
Inspired by Matlab plotting
Both Matlab-like & OOP syntax
Abstracts many underlying details
Usually better default aesthetics
Relation to Pandas
MATPLOTLIB
SEABORN
There is a deep integration of these packages with Pandas
As we go through this session,
we will learn new ideas in Pandas
PANDAS
FDS_Visualisation
By One Fourth Labs
FDS_Visualisation
Visualisation
