1
00:00:00,000 --> 00:00:03,694
[MUSIC]

2
00:00:03,694 --> 00:00:04,694
Hi, everyone.

3
00:00:04,694 --> 00:00:07,151
In this video we'll learn
how to use Kaggle for

4
00:00:07,151 --> 00:00:10,160
participation in data
science competitions.

5
00:00:10,160 --> 00:00:11,361
Let's open kaggle.com.

6
00:00:11,361 --> 00:00:16,800
On the Competitions page, we can see
a list of currently running competitions.

7
00:00:16,800 --> 00:00:21,446
Every competition has a page which
consists of title, short description,

8
00:00:21,446 --> 00:00:25,958
price budget, number of participating
teams, and time before the end.

9
00:00:25,958 --> 00:00:29,869
Information involves all
previously running competitions,

10
00:00:29,869 --> 00:00:31,719
we can find if we click to All.

11
00:00:31,719 --> 00:00:36,137
Let's select some challenge and
see how it organized.

12
00:00:39,777 --> 00:00:45,240
Here, we see several tabs which we'll
explore, and let's start with Overview.

13
00:00:46,990 --> 00:00:51,790
In the Description section we see
an introduction provided by organizers.

14
00:00:51,790 --> 00:00:55,006
In the Description, there is a short
story about company and tasks,

15
00:00:55,006 --> 00:00:56,515
sometimes with illustration.

16
00:00:56,515 --> 00:01:02,160
At the Evaluation page, we see
the description of the target metric.

17
00:01:02,160 --> 00:01:05,786
In this challenge, target metric
is the Mean Absolute Error between

18
00:01:05,786 --> 00:01:10,080
the logarithmic transform predictions and
ground truth values.

19
00:01:10,080 --> 00:01:14,708
This page also contains example of sample
submission file, which is typical for

20
00:01:14,708 --> 00:01:16,350
such kind of competitions.

21
00:01:17,730 --> 00:01:19,230
Now let's move to the Prize page.

22
00:01:20,750 --> 00:01:24,070
In the Prize,
page we can find information about prizes.

23
00:01:24,070 --> 00:01:29,832
Take notice that in the title we have
information about the whole money budget,

24
00:01:29,832 --> 00:01:33,970
and this page,
we see how it will be split among winners.

25
00:01:35,110 --> 00:01:38,420
I want to highlight that
in order to get money,

26
00:01:38,420 --> 00:01:43,600
you need not only be in top three teams,
but also beat a Zillow benchmark model.

27
00:01:45,300 --> 00:01:52,140
Now let's see, Timeline page, which
contains all the information about dates.

28
00:01:52,140 --> 00:01:55,113
For example,
when competition starts, ends,

29
00:01:55,113 --> 00:01:58,530
when will the Team Merger deadline and
then what month.

30
00:02:00,020 --> 00:02:03,870
All the details about competition,
we can find in the Rules.

31
00:02:03,870 --> 00:02:06,621
So we need to check really the rules.

32
00:02:06,621 --> 00:02:11,147
Here we can find that team
limit is three individual,

33
00:02:11,147 --> 00:02:16,276
that we have maximum of five
submissions per day, that you,

34
00:02:16,276 --> 00:02:21,726
for example, should be at least
18 years old to participate.

35
00:02:21,726 --> 00:02:24,695
And that, find it,

36
00:02:24,695 --> 00:02:30,123
that external data are not allowed.

37
00:02:30,123 --> 00:02:35,540
I strongly suggest you to read the rules
carefully before joining the competition.

38
00:02:35,540 --> 00:02:38,560
And after reading, you should accept it,
but I already accepted it.

39
00:02:41,440 --> 00:02:43,320
Now, let's check this, Data.

40
00:02:45,700 --> 00:02:49,940
Here we have data provided by
the organizers, several files which we can

41
00:02:49,940 --> 00:02:55,680
download, and sample submission among
them, and the description of the data.

42
00:02:55,680 --> 00:03:01,070
Here we have description of files,
description of data fields, and

43
00:03:01,070 --> 00:03:04,250
more importantly a description
of train and test split.

44
00:03:04,250 --> 00:03:09,190
This is quite useful information in
order to set up right validation scheme.

45
00:03:10,760 --> 00:03:15,340
If you have any question about data or
other questions to ask, or insights to

46
00:03:15,340 --> 00:03:18,770
share, you can go to the forum,
which we can find under Discussion tab.

47
00:03:20,760 --> 00:03:25,446
Usually it contain a lot of topics or
threads, like Welcome, questions about

48
00:03:25,446 --> 00:03:29,790
validations, questions about train and
test data, and so on and so on.

49
00:03:31,200 --> 00:03:36,730
Every topic have title,
number of comments, and number of reports.

50
00:03:36,730 --> 00:03:38,618
Let's see some of them.

51
00:03:38,618 --> 00:03:41,670
Here we have main message,
a lot of comments,

52
00:03:41,670 --> 00:03:45,170
in this particular we
have only one comments.

53
00:03:45,170 --> 00:03:54,230
Each we can up vote or down vote and
reply to by click the reply button.

54
00:03:55,910 --> 00:03:59,900
That was a brief overview on forum and
now we switch to the Kernels.

55
00:03:59,900 --> 00:04:04,160
Usually, I run my code locally,
but sometimes it would be handy

56
00:04:04,160 --> 00:04:09,200
to check an idea quickly or share code
with other participants or teammates.

57
00:04:09,200 --> 00:04:11,270
This is what Kernels are for.

58
00:04:11,270 --> 00:04:16,080
You can think of Kernel as a small virtual
machine in which you write your code,

59
00:04:16,080 --> 00:04:18,320
execute it, and share it.

60
00:04:18,320 --> 00:04:21,374
Let's take a look at some Kernel,
for example for this one.

61
00:04:25,140 --> 00:04:30,130
This show explanatory data analysis
on the Zillow competition.

62
00:04:30,130 --> 00:04:35,960
It took quite long, contain a lot of
pictures, and I believe it very useful.

63
00:04:37,290 --> 00:04:40,425
Here we can see comments for
this, different versions.

64
00:04:40,425 --> 00:04:44,881
And in order,
if you want to make a copy and edit it,

65
00:04:44,881 --> 00:04:47,586
we need to Fork this Notebook.

66
00:04:47,586 --> 00:04:52,231
It doesn't matter how your
predictions were produced, locally or

67
00:04:52,231 --> 00:04:57,210
by Kernel, you should submit
them through a specialized form.

68
00:04:57,210 --> 00:04:58,611
So go back to the competition.

69
00:05:00,634 --> 00:05:01,900
Go to submissions.

70
00:05:01,900 --> 00:05:07,050
I already submit sample submission,
you can do the same.

71
00:05:07,050 --> 00:05:11,290
Click submit predictions,
and drag and drop file here.

72
00:05:11,290 --> 00:05:12,048
Let's look at my submission.

73
00:05:17,680 --> 00:05:19,905
After submission,
you will see it on the leaderboard.

74
00:05:20,950 --> 00:05:23,260
This is my sample submission.

75
00:05:23,260 --> 00:05:27,470
Leaderboard contains information
about all the teams.

76
00:05:27,470 --> 00:05:34,270
So here we have team name or just name
in case of single competition team.

77
00:05:34,270 --> 00:05:38,462
Score which we produced,
number of submissions,

78
00:05:38,462 --> 00:05:45,070
time since the last submissions, and
position data over seven last days.

79
00:05:45,070 --> 00:05:51,940
For example, this means that this guy
drops 19 positions during the last week.

80
00:05:53,430 --> 00:05:56,220
That was a brief overview
of Kaggle interface.

81
00:05:56,220 --> 00:05:59,160
Further, I will tell some extra
information about the platform.

82
00:06:00,270 --> 00:06:04,770
So let's move to Overview
page at the bottom.

83
00:06:04,770 --> 00:06:09,410
And here,
we see information about points and tiers.

84
00:06:09,410 --> 00:06:14,790
As mentioned here, the competition will be
counting towards ranking points an tiers.

85
00:06:14,790 --> 00:06:17,820
If you participate,
it will be beneficial for your rating.

86
00:06:18,910 --> 00:06:22,780
Sometimes, especially in educational
competitions, it's not like that.

87
00:06:24,950 --> 00:06:29,930
Information about Kaggle Progression
System we can find if we click this link,

88
00:06:32,050 --> 00:06:36,780
where we can read information about
tiers like novice, contributor, master,

89
00:06:36,780 --> 00:06:37,522
grandmaster.

90
00:06:37,522 --> 00:06:40,066
About medals and ranking points.

91
00:06:41,721 --> 00:06:45,827
This ranking points, I use for
global User Ranking.

92
00:06:45,827 --> 00:06:47,680
Let's check it.

93
00:06:51,540 --> 00:06:54,920
So, we have user ranking page, and

94
00:06:54,920 --> 00:06:59,370
we see all the users ranked, and
with links to their profile.

95
00:06:59,370 --> 00:07:01,471
Let's check some profile,
for example mine.

96
00:07:01,471 --> 00:07:06,911
And here we have photo,
name, some information,

97
00:07:06,911 --> 00:07:15,400
geo information, information about
past competitions, medals, and so on.

98
00:07:15,400 --> 00:07:20,080
In addition, I want to say a few words
about ability to host competition.

99
00:07:21,080 --> 00:07:21,889
Kaggle has this ability.

100
00:07:21,889 --> 00:07:28,020
Click Host competition, and
there is special Kaggle in class.

101
00:07:29,645 --> 00:07:33,470
At in class, everyone can host
their own competition for free and

102
00:07:33,470 --> 00:07:35,530
invite people to participate.

103
00:07:35,530 --> 00:07:38,990
This option is quite often used in
various educational competitions.

104
00:07:40,160 --> 00:07:42,394
So this was a brief overview
of Kaggle platform.

105
00:07:42,394 --> 00:07:43,814
Thank for your attention.

106
00:07:43,814 --> 00:07:53,814
[MUSIC]