1
00:00:08,650 --> 00:00:12,616
Welcome to the second course in the
Introduction to Applied Data Science with

2
00:00:12,616 --> 00:00:14,840
Python specialization.

3
00:00:14,840 --> 00:00:18,060
In the first course we looked,
primarily, at how to manipulate data,

4
00:00:18,060 --> 00:00:20,390
including loading data from
structured delimited files,

5
00:00:20,390 --> 00:00:25,160
such as CSVs, cleaning the data with
Python pandas, and NumPy libraries, and

6
00:00:25,160 --> 00:00:29,770
transforming data using functions,
such as apply, aggregate, map, and pivot.

7
00:00:29,770 --> 00:00:34,180
If you don't feel you have a solid
understanding of the pandas toolkit,

8
00:00:34,180 --> 00:00:37,330
then I would invite you to go back and
review that course content.

9
00:00:39,450 --> 00:00:43,550
The main topic of this course is
understanding how to communicate data more

10
00:00:43,550 --> 00:00:47,990
clearly to the various stakeholders
who are interested in the data.

11
00:00:47,990 --> 00:00:51,150
And there are many different
stakeholders to whom you might be

12
00:00:51,150 --> 00:00:53,050
interested in using visuals to reach.

13
00:00:54,050 --> 00:00:57,910
For instance, you want to share unbiased,
representations of data with

14
00:00:57,910 --> 00:01:02,860
peer data scientists who are able to make
their own inferences and interpretations.

15
00:01:02,860 --> 00:01:06,560
But you'll also want to be able to share
stories about the data with managers and

16
00:01:06,560 --> 00:01:09,820
executives who might be familiar
with domain but are looking for

17
00:01:09,820 --> 00:01:11,740
specific actionable insights.

18
00:01:13,040 --> 00:01:16,744
Finally, you'll also want to be able to
communicate like a journalist to a broad

19
00:01:16,744 --> 00:01:19,733
population who knows very little
about your particular data and

20
00:01:19,733 --> 00:01:23,496
needs to understand not only an overview
of the phenomenon you are describing but

21
00:01:23,496 --> 00:01:26,061
also the limitations of
the approaches you have used.

22
00:01:27,936 --> 00:01:32,590
This course is about specific techniques
which are part of a broad subject of

23
00:01:32,590 --> 00:01:34,490
information visualization.

24
00:01:35,600 --> 00:01:37,800
Referred to an infovis for short,

25
00:01:37,800 --> 00:01:42,660
this is a field that has rapidly developed
over the past 20 years, as computing power

26
00:01:42,660 --> 00:01:47,180
has made it possible to build compelling
visual and interactive graphics quickly.

27
00:01:48,210 --> 00:01:51,990
And while the use of visualizations to
impart knowledge of data has been around

28
00:01:51,990 --> 00:01:56,840
for centuries, the last 20 years have seen
the development of important theoretical

29
00:01:56,840 --> 00:02:02,330
basis for best practices in displaying
information to various audiences.

30
00:02:02,330 --> 00:02:04,770
And in this course,
we're going to focus on those.

31
00:02:06,570 --> 00:02:10,020
In this first week of the course, we're
going to touch on some of those theories

32
00:02:10,020 --> 00:02:14,050
and best practices, and talk about
fundamental thinkers in the field.

33
00:02:14,050 --> 00:02:17,537
In particular,
Alberto Cairo and Edward Tufte.

34
00:02:17,537 --> 00:02:19,999
We'll go through some examples
of excellent and not so

35
00:02:19,999 --> 00:02:21,347
excellent visualizations.

36
00:02:21,347 --> 00:02:22,999
And develop a language for

37
00:02:22,999 --> 00:02:27,660
speaking about how we might make
meaningful visualizations ourselves.

38
00:02:28,790 --> 00:02:33,060
This portion of the course will be largely
lecture based with several readings and

39
00:02:33,060 --> 00:02:35,170
peer reviewed assignment
at the end of each week.

40
00:02:36,670 --> 00:02:39,920
In the second and third week of
the course, we'll dive into the de facto

41
00:02:39,920 --> 00:02:44,440
Python library for creating charts and
graphs, matplotlib.

42
00:02:44,440 --> 00:02:48,670
Now matplotlib isn't the only way to
create visual artifacts in Python, but

43
00:02:48,670 --> 00:02:53,870
it's the basis for some emerging new
web tools such as Seaborn and Bokeh.

44
00:02:53,870 --> 00:02:56,470
And acts as a solid foundation for
our discussions.

45
00:02:57,510 --> 00:03:01,750
The majority of these weeks will be spent
actually working on assignments with

46
00:03:01,750 --> 00:03:05,430
a few lectures to share an overview
of how the toolkit itself works.

47
00:03:06,480 --> 00:03:10,530
And you'll be expected to take messy
raw data, process it using pandas or

48
00:03:10,530 --> 00:03:15,320
other suitable techniques and render it
into a meaningful, explanatory image.

49
00:03:16,400 --> 00:03:20,340
And again, if you're not comfortable yet
using pandas, I would encourage you to go

50
00:03:20,340 --> 00:03:25,060
back to the first course in this
specialization and review those videos.

51
00:03:25,060 --> 00:03:28,390
The assignments for weeks two and
three will be largely peer graded.

52
00:03:29,990 --> 00:03:33,190
Finally, the course will
culminate with a small project,

53
00:03:33,190 --> 00:03:35,320
similar to the first course.

54
00:03:35,320 --> 00:03:38,550
In this project, you will be given
a new visualization technique and

55
00:03:38,550 --> 00:03:43,540
you will be asked to extend the matplotlib
code base to enhance its functionality.

56
00:03:43,540 --> 00:03:46,910
This is a critical step in your
journey as a data scientist.

57
00:03:46,910 --> 00:03:50,640
To go beyond using tools off the shelf and
instead be able to modify

58
00:03:50,640 --> 00:03:53,780
them to support new ways of
understanding data and problems.

59
00:03:55,630 --> 00:03:57,620
This course is very similar to the first,

60
00:03:57,620 --> 00:03:59,790
except the evaluation
is largely peer review.

61
00:03:59,790 --> 00:04:03,650
The lectures will only guide you so
much in completing the assignments.

62
00:04:03,650 --> 00:04:07,500
You'll need to know how to ask questions
in the discussion forums of your peers.

63
00:04:07,500 --> 00:04:11,560
And seek out new information through
web searches and Stack Overflow.

64
00:04:11,560 --> 00:04:14,150
And again,
Stack Overflow is tagged specifically for

65
00:04:14,150 --> 00:04:16,150
questions about matplotlib.

66
00:04:16,150 --> 00:04:18,530
So I'd encourage you to
use that resource as well.

67
00:04:19,590 --> 00:04:20,710
Finally, for this course,

68
00:04:20,710 --> 00:04:23,740
I expect that you are familiar with
all of the content from Course 1.

69
00:04:23,740 --> 00:04:27,780
I'm looking forward to this
next course in specialization.

70
00:04:27,780 --> 00:04:28,400
Let's get started.