1
00:00:00,000 --> 00:00:09,320
[MUSIC]

2
00:00:09,320 --> 00:00:11,260
Hello, I'm Craig, and I'm a developer.

3
00:00:11,260 --> 00:00:12,053
In this course,

4
00:00:12,053 --> 00:00:15,404
we're going to to be exploring
the wonderful data library, pandas.

5
00:00:15,404 --> 00:00:19,013
Now, pandas is a portmanteau,
or a combination of two words,

6
00:00:19,013 --> 00:00:21,337
in this case, the words panel and data.

7
00:00:21,337 --> 00:00:23,685
Panel data is data that
is multidimensional,

8
00:00:23,685 --> 00:00:25,556
involving measurements over time.

9
00:00:25,556 --> 00:00:29,567
pandas are also an adorable creature, and
I hope that you're here for the former,

10
00:00:29,567 --> 00:00:33,260
but I totally understand that I might
have clickbaited you into the latter.

11
00:00:33,260 --> 00:00:37,487
pandas provides fast, flexible, and
expressive data structures that have been

12
00:00:37,487 --> 00:00:39,753
designed to make working
with relational or

13
00:00:39,753 --> 00:00:42,342
labeled data not only easy,
but also intuitive.

14
00:00:42,342 --> 00:00:46,186
It's the fundamental high level
building block for doing practical and

15
00:00:46,186 --> 00:00:48,148
real world data analysis in Python.

16
00:00:48,148 --> 00:00:51,267
Before we get cooking, let's make
sure that we're on the same page.

17
00:00:51,267 --> 00:00:54,321
There are definitely some
prerequisites for this course,

18
00:00:54,321 --> 00:00:56,956
so please double check
that you're all caught up.

19
00:00:56,956 --> 00:00:59,321
The most important of
the prerequisites is NumPy.

20
00:00:59,321 --> 00:01:03,887
I'd like to make sure that you had a nice
introduction to the NumPy library.

21
00:01:03,887 --> 00:01:05,797
pandas relies heavily on NumPy, and

22
00:01:05,797 --> 00:01:10,305
I'm going to assume that you have a basic
understanding of its overarching concepts.

23
00:01:10,305 --> 00:01:13,047
Now, don't worry if it's been
a while since you've used it,

24
00:01:13,047 --> 00:01:15,920
we'll retouch upon the concepts
that you need here in just a bit.

25
00:01:15,920 --> 00:01:19,468
Don't forget to check the teacher's
notes that are attached to each video.

26
00:01:19,468 --> 00:01:21,346
I'll try and
remind you to look in there, but

27
00:01:21,346 --> 00:01:23,677
please do get in the habit of
checking that section out.

28
00:01:23,677 --> 00:01:27,130
Lots of great information is tucked away
in there waiting for you to dig into it.

29
00:01:27,130 --> 00:01:30,091
In this course,
I'm gonna try a new approach.

30
00:01:30,091 --> 00:01:34,043
In an effort to give you more practice
of how data professionals interact,

31
00:01:34,043 --> 00:01:37,510
I'm going to rely more heavily
than usual on Jupyter notebooks.

32
00:01:37,510 --> 00:01:39,326
As you are most likely already aware,

33
00:01:39,326 --> 00:01:42,407
Jupyter notebooks are a great
place to capture your learnings.

34
00:01:42,407 --> 00:01:45,212
They're also intended to be used for
teaching.

35
00:01:45,212 --> 00:01:48,768
I've gone ahead and build up some
interactive content that will assist you

36
00:01:48,768 --> 00:01:50,411
in exploring the pandas library.

37
00:01:50,411 --> 00:01:51,427
In the Treehouse app,

38
00:01:51,427 --> 00:01:54,486
you'll encounter these notebooks
as textual instruction steps.

39
00:01:54,486 --> 00:01:57,913
I've included information in the teacher's
notes about how to get a hold of

40
00:01:57,913 --> 00:02:00,818
the notebooks so that you can run them and
follow along locally.

41
00:02:00,818 --> 00:02:02,850
I'd love for you, as a lifelong learner,

42
00:02:02,850 --> 00:02:06,129
to get in the habit of exploring
every notebook that you come across.

43
00:02:06,129 --> 00:02:08,540
Use it to poke around as
you learn a new library,

44
00:02:08,540 --> 00:02:11,148
much like you might expect
to use the Python shell.

45
00:02:11,148 --> 00:02:15,386
Explore the API and practice different
approaches, and most importantly,

46
00:02:15,386 --> 00:02:16,570
keep your own notes.

47
00:02:16,570 --> 00:02:19,884
A common data science workflow
involves multiple stages.

48
00:02:19,884 --> 00:02:22,322
First you clean the data and
then you analyze and model it.

49
00:02:22,322 --> 00:02:26,520
And finally, you organize the results
of the analysis into either a graph or

50
00:02:26,520 --> 00:02:27,052
a table.

51
00:02:27,052 --> 00:02:30,194
Great news, pandas can do all that,
the entire workflow.

52
00:02:30,194 --> 00:02:34,165
Even better news,
it's really a pleasure to use.

53
00:02:34,165 --> 00:02:38,527
Since you already have a fundamental
understanding of the numerical library,

54
00:02:38,527 --> 00:02:41,467
NumPy, pandas is going to
feel very familiar to you.

55
00:02:41,467 --> 00:02:45,455
In fact, pandas sits directly on
top of NumPy like a little hat.

56
00:02:45,455 --> 00:02:46,703
I don't know about you, but

57
00:02:46,703 --> 00:02:49,875
one of the things that I have trouble
with in NumPy is when I have an array.

58
00:02:49,875 --> 00:02:53,183
I never know just which value is which.

59
00:02:53,183 --> 00:02:58,194
Like for instance, in this array here, I
don't really know who got the high score.

60
00:02:58,194 --> 00:03:02,091
I have to remember that Robbie is
the first one here at index zero.

61
00:03:02,091 --> 00:03:03,416
But I just have to know that.

62
00:03:03,416 --> 00:03:07,023
pandas gives you a new ability,
you can label each value.

63
00:03:07,023 --> 00:03:09,484
It's like a dictionary, a key and a value.

64
00:03:09,484 --> 00:03:11,786
And that works great for
a single dimension.

65
00:03:11,786 --> 00:03:16,238
This example is the series of high
scorers for a single game, Donkey Kong,

66
00:03:16,238 --> 00:03:18,227
labeled by players' initials.

67
00:03:18,227 --> 00:03:21,757
But as you know, we often want
to have multidimensional data.

68
00:03:21,757 --> 00:03:25,023
We could track more games by
adding a new game dimension,

69
00:03:25,023 --> 00:03:27,076
like we could add Pac-Man scores.

70
00:03:27,076 --> 00:03:29,693
But now we have to remember two indexes,
and

71
00:03:29,693 --> 00:03:34,164
I have to remember that index zero is
Donkey Kong and index one is Pac-Man.

72
00:03:34,164 --> 00:03:37,641
Now, again,
pandas does a great job with labeling.

73
00:03:37,641 --> 00:03:41,629
You can also label each of these columns,
so you end up with tabular data.

74
00:03:41,629 --> 00:03:46,010
The two-dimensioned data structure
here is known as a data frame.

75
00:03:46,010 --> 00:03:49,698
This is a data frame of high scores
on multiple games indexed by players'

76
00:03:49,698 --> 00:03:50,307
initials.

77
00:03:50,307 --> 00:03:54,555
And that ought to feel pretty familiar,
assuming you've used tabular or

78
00:03:54,555 --> 00:03:58,392
table based data before,
like a spreadsheet or a database table,

79
00:03:58,392 --> 00:04:00,325
anything with rows and columns.

80
00:04:00,325 --> 00:04:03,276
With pandas,
you can put any sort of data in there too.

81
00:04:03,276 --> 00:04:05,869
It doesn't have the same
restrictions like NumPy did.

82
00:04:05,869 --> 00:04:08,835
pandas also lets you
relate datasets by label.

83
00:04:08,835 --> 00:04:09,838
So you can merge and

84
00:04:09,838 --> 00:04:13,611
join together related information
in a very straightforward manner.

85
00:04:13,611 --> 00:04:18,167
pandas is a full-featured library, and we
simply won't be able to get to all of its

86
00:04:18,167 --> 00:04:20,715
amazing powers in this
introductory course.

87
00:04:20,715 --> 00:04:24,476
I do hope to give you a firm foundation
and guide you to where you can learn and

88
00:04:24,476 --> 00:04:25,324
practice more.

89
00:04:25,324 --> 00:04:29,088
For this course, I'm gonna ask that you
imagine that there is a new company

90
00:04:29,088 --> 00:04:32,931
in town jumping in on that social banking
app craze, like Cash App or Venmo.

91
00:04:32,931 --> 00:04:34,778
They call themselves Cash Box.

92
00:04:34,778 --> 00:04:39,138
Basically the way that their app works is
that a user signs up, chooses a username,

93
00:04:39,138 --> 00:04:43,079
and then they can send money to other
users of the system by their username.

94
00:04:43,079 --> 00:04:46,157
Now, a common use case for
their app is when it's lunch time and

95
00:04:46,157 --> 00:04:47,794
people don't have cash on them.

96
00:04:47,794 --> 00:04:51,225
Their users can just send money through
Cash Box to the person picking up

97
00:04:51,225 --> 00:04:51,748
the bill.

98
00:04:51,748 --> 00:04:55,284
Now, each user on Cash Box keeps
a balance of their funds, and

99
00:04:55,284 --> 00:04:57,394
the app tracks their transactions.

100
00:04:57,394 --> 00:05:03,188
Good news, Cash Box is hiring and they
are looking for a junior data scientist.

101
00:05:03,188 --> 00:05:05,064
They've sent out a hiring challenge and

102
00:05:05,064 --> 00:05:07,170
access to a sample of
some of their datasets.

103
00:05:07,170 --> 00:05:09,888
So what do you say we
explore their data sets and

104
00:05:09,888 --> 00:05:12,127
pick up some job skills along the way?

105
00:05:12,127 --> 00:05:14,090
Let's get ready to rock the Cash Box.

106
00:05:14,090 --> 00:05:18,737
[LAUGH] Good thing we aren't applying
to be part of the marketing department.