1
00:00:00,000 --> 00:00:02,010
Hi everyone.

2
00:00:02,010 --> 00:00:04,065
This is the presentation for

3
00:00:04,065 --> 00:00:06,840
the paper of Thirty-two Years of IEEE VIS:

4
00:00:06,840 --> 00:00:11,415
Authors, Fields of Study and Citations.

5
00:00:11,415 --> 00:00:14,880
VIS recently positioned itself 

6
00:00:14,880 --> 00:00:16,785
within the field
of data science.

7
00:00:16,785 --> 00:00:19,140
However, this does not tell us

8
00:00:19,140 --> 00:00:21,810
where VIS is in
science overall.

9
00:00:21,810 --> 00:00:24,555
For example, we don't
know which fields

10
00:00:24,555 --> 00:00:30,550
VIS is citing and which
fields are citing VIS.

11
00:00:30,640 --> 00:00:33,770
The official website of 

12
00:00:33,770 --> 00:00:36,710
IEEE VIS mentioned
that: The conference will

13
00:00:36,710 --> 00:00:39,290
convene an international
community of

14
00:00:39,290 --> 00:00:42,470
researchers and practitioners
from universities,

15
00:00:42,470 --> 00:00:46,355
government, and industry to
exchange recent findings.

16
00:00:46,355 --> 00:00:48,650
However, we don't know

17
00:00:48,650 --> 00:00:50,210
how many authors are from

18
00:00:50,210 --> 00:00:53,090
universities, government,
and industry.

19
00:00:53,090 --> 00:00:55,130
That is, we don't know

20
00:00:55,130 --> 00:00:58,385
the statistics of
author affiliations

21
00:00:58,385 --> 00:01:00,920
in the past 32 years.

22
00:01:00,920 --> 00:01:04,640
Here, we try to answer
these two questions.

23
00:01:04,640 --> 00:01:08,000
First, where VIS stands in science?

24
00:01:08,000 --> 00:01:10,460
And second, where VIS authors

25
00:01:10,460 --> 00:01:13,770
are from and how
they collaborated.

26
00:01:13,890 --> 00:01:16,750
To answer these two questions,

27
00:01:16,750 --> 00:01:20,320
we collected all the
relevant paper DOIs

28
00:01:20,320 --> 00:01:28,435
in the past 32 years
from 1990 to 2021.

29
00:01:28,435 --> 00:01:32,740
Based on those DOIs, we
collected data on paper,

30
00:01:32,740 --> 00:01:36,850
authors and fields of study
from OpenAlex and IEEEXplore

31
00:01:36,850 --> 00:01:40,105
More details 

32
00:01:40,105 --> 00:01:42,910
about the procedure can
be found in this diagram.

33
00:01:42,910 --> 00:01:45,430
And we published our dataset

34
00:01:45,430 --> 00:01:48,775
in the official website
of our project.

35
00:01:48,775 --> 00:01:51,430
Here we'll report the results.

36
00:01:51,430 --> 00:01:53,665
In terms of the general trends.

37
00:01:53,665 --> 00:01:57,755
We found that VIS has
been increasingly popular,

38
00:01:57,755 --> 00:02:00,380
which is evident

39
00:02:00,380 --> 00:02:02,360
in the increasing number of

40
00:02:02,360 --> 00:02:04,340
publications each year and

41
00:02:04,340 --> 00:02:07,475
the increasing number of
unique authors each year.

42
00:02:07,475 --> 00:02:12,005
VIS is also becoming
more impactful because

43
00:02:12,005 --> 00:02:13,280
an increasing number of

44
00:02:13,280 --> 00:02:17,000
citations are from
non-VIS papers.

45
00:02:17,000 --> 00:02:19,010
We found that there were

46
00:02:19,010 --> 00:02:22,085
more and more collaborations
in VIS because

47
00:02:22,085 --> 00:02:24,440
the proportion of
cross-country and

48
00:02:24,440 --> 00:02:27,335
cross-type collaborations
has been increasing.

49
00:02:27,335 --> 00:02:30,560
By "cross-type", we mean
the collaboration between

50
00:02:30,560 --> 00:02:32,390
authors from universities

51
00:02:32,390 --> 00:02:35,390
are non-educational
affiliations.

52
00:02:35,390 --> 00:02:38,435
Those collaborations,
however, were

53
00:02:38,435 --> 00:02:41,420
concentrated because,
for example,

54
00:02:41,420 --> 00:02:46,640
the graph here shows that
the top ten countries were

55
00:02:46,640 --> 00:02:49,010
present in 98% of

56
00:02:49,010 --> 00:02:51,200
all the cross-country
collaborations

57
00:02:51,200 --> 00:02:53,790
in the past 32 years.

58
00:02:53,950 --> 00:02:58,985
In terms of the geographical
aspect of authors,

59
00:02:58,985 --> 00:03:02,945
we've found that an
increasing number

60
00:03:02,945 --> 00:03:05,630
of countries are
participating in VIS.

61
00:03:05,630 --> 00:03:08,435
For example, in 2021,

62
00:03:08,435 --> 00:03:12,140
authors from 26 countries
participated in VIS,

63
00:03:12,140 --> 00:03:13,820
whereas only five countries

64
00:03:13,820 --> 00:03:18,740
participating in the first
VIS conference in 1990.

65
00:03:18,740 --> 00:03:25,865
This participatino is
concentrated in terms of 

66
00:03:25,865 --> 00:03:30,185
author country origins and also
in terms of the continent.

67
00:03:30,185 --> 00:03:34,940
There are some redistributions
in the participation.

68
00:03:34,940 --> 00:03:37,040
For example, the percentage of

69
00:03:37,040 --> 00:03:39,800
authors from the
United States has been

70
00:03:39,800 --> 00:03:43,475
constantly declining
and the percentage

71
00:03:43,475 --> 00:03:46,830
of authors from China
has been increasing.

72
00:03:47,590 --> 00:03:50,765
In terms of author
affiliation types.

73
00:03:50,765 --> 00:03:55,640
We've found that authors from
universities dominated VIS.

74
00:03:55,640 --> 00:03:57,035
Earlier.

75
00:03:57,035 --> 00:04:00,275
We showed that cross-type collaborations

76
00:04:00,275 --> 00:04:02,120
have been increasing.

77
00:04:02,120 --> 00:04:06,560
However, here we've found that the
proportion of authors from

78
00:04:06,560 --> 00:04:08,480
non-educational affiliations

79
00:04:08,480 --> 00:04:11,015
has been constantly declining.

80
00:04:11,015 --> 00:04:14,285
The right panel shows
that although there is

81
00:04:14,285 --> 00:04:15,500
an increasing number of

82
00:04:15,500 --> 00:04:18,785
authors from educational
affiliations,

83
00:04:18,785 --> 00:04:20,480
the number of authors from

84
00:04:20,480 --> 00:04:22,640
non-educational affiliations,

85
00:04:22,640 --> 00:04:29,855
has been stabilizing at around
a 100 the past 32 years.

86
00:04:29,855 --> 00:04:32,150
So combining these two results,

87
00:04:32,150 --> 00:04:33,620
we can know that

88
00:04:33,620 --> 00:04:36,290
the small but stable number

89
00:04:36,290 --> 00:04:38,225
of authors from non-educational

90
00:04:38,225 --> 00:04:40,940
affiliations have been actively

91
00:04:40,940 --> 00:04:45,740
participating in VIS
projects in the past 32 years.

92
00:04:45,740 --> 00:04:48,350
In terms of fields of study,

93
00:04:48,350 --> 00:04:52,280
this graph shows
the distribution of

94
00:04:52,280 --> 00:04:54,845
the lowest level concepts or

95
00:04:54,845 --> 00:04:57,575
fields of study in VIS papers.

96
00:04:57,575 --> 00:04:59,960
And we've found that most
of the papers were about

97
00:04:59,960 --> 00:05:02,465
computer science
and mathematics.

98
00:05:02,465 --> 00:05:07,160
The same is for papers that
are referenced in VIS.

99
00:05:07,160 --> 00:05:09,275
And papers that are citing VIS.

100
00:05:09,275 --> 00:05:13,145
And that's why we say
that VIS is mainly

101
00:05:13,145 --> 00:05:15,425
about, built upon, and

102
00:05:15,425 --> 00:05:19,325
impacting computer
science and mathematics.

103
00:05:19,325 --> 00:05:22,595
We also found a concentration

104
00:05:22,595 --> 00:05:25,400
of concepts in these papers.

105
00:05:25,400 --> 00:05:29,480
Because at each level,

106
00:05:29,480 --> 00:05:32,240
only a few concepts were

107
00:05:32,240 --> 00:05:35,990
frequently appearing

108
00:05:35,990 --> 00:05:38,030
in VIS papers.

109
00:05:38,030 --> 00:05:41,480
We also examined the citation

110
00:05:41,480 --> 00:05:44,840
flows based on fields of study.

111
00:05:44,840 --> 00:05:47,870
The left panel shows
the citation flows

112
00:05:47,870 --> 00:05:53,360
from papers that are
referenced in VIS to VIS.

113
00:05:53,360 --> 00:05:55,220
And the right panel shows

114
00:05:55,220 --> 00:05:57,500
the citations flows
from VIS papers,

115
00:05:57,500 --> 00:05:59,930
to papers that
are citing VIS.

116
00:05:59,930 --> 00:06:02,645
And these two figures showed

117
00:06:02,645 --> 00:06:05,420
that citations mostly flow

118
00:06:05,420 --> 00:06:08,225
between the same
set of concepts.

119
00:06:08,225 --> 00:06:11,870
The interactive
visualizations can be

120
00:06:11,870 --> 00:06:16,980
found in the official
website of our project.

121
00:06:17,350 --> 00:06:20,360
In terms of citations,

122
00:06:20,360 --> 00:06:22,790
we found that the
lion's share of

123
00:06:22,790 --> 00:06:26,060
citations were taken
by the top papers.

124
00:06:26,060 --> 00:06:31,700
For example,

125
00:06:31,700 --> 00:06:33,605
the top 20% papers,

126
00:06:33,605 --> 00:06:38,360
took 60% of all the citations.

127
00:06:38,360 --> 00:06:42,710
We ran a regression analysis to

128
00:06:42,710 --> 00:06:47,165
see what factors are influencing
the number of citations.

129
00:06:47,165 --> 00:06:51,155
And we found that earlier
works, journal papers,

130
00:06:51,155 --> 00:06:53,435
and papers that have won

131
00:06:53,435 --> 00:06:57,950
an award have significantly
more citations.

132
00:06:57,950 --> 00:07:01,760
To recap, we've found
that VIS has been

133
00:07:01,760 --> 00:07:04,385
becoming increasingly popular,

134
00:07:04,385 --> 00:07:06,995
impactful and collaborative.

135
00:07:06,995 --> 00:07:11,390
We've found that geographically,
authors are diverse but

136
00:07:11,390 --> 00:07:13,625
concentrated. In terms of

137
00:07:13,625 --> 00:07:15,740
author affiliation
types, we found

138
00:07:15,740 --> 00:07:19,295
that authors from
universities dominated VIS.

139
00:07:19,295 --> 00:07:21,290
In terms of fields of study,

140
00:07:21,290 --> 00:07:24,470
we found that VIS is
mainly about, built upon

141
00:07:24,470 --> 00:07:28,160
and impacting computer
science and mathematics.

142
00:07:28,160 --> 00:07:30,230
And we also found about

143
00:07:30,230 --> 00:07:34,940
citations flow mostly between
the same set of concepts.

144
00:07:34,940 --> 00:07:37,370
Our regression analysis

145
00:07:37,370 --> 00:07:38,930
shows that earlier works,

146
00:07:38,930 --> 00:07:41,255
journal papers and award

147
00:07:41,255 --> 00:07:44,825
winning papers had
more citations.

148
00:07:44,825 --> 00:07:49,470
Thank you, and I'm happy
to take questions.