Methodology: In the context of this work, we define superspreaders as individuals that have a large number of contacts within the course of a time interval. To assess the impact of these individuals on the contact network, we construct a graph for each 5-minute interval, then we identify the top N people that had the largest number of contacts within this specific time interval and then we remove them from the graph. Our analysis assess the percentage of contacts that remain in the contact networks after removing the top 2, 3, 4, 5, 6, 8, 10, 15, 20, and 50 superspreaders. Below we report our results and the main take away points from our analysis.

Main Insights: Our experimental results indicate that indeed a small percentage of people are responsible for a large number of contacts in our contact network. For instance, by looking at the last column in Table 1, by removing the top 10 superspreaders (approx. 1% of the entire population), we effectively remove over 26% of all the contacts.
Figure 1a: Cumulative Distribution Function of the number of average contacts for each observation before and after removing top N superspreaders.
Figure 1b: Cumulative Distribution Function of the percentage of contacts that remain after removing top N superspreaders for each observation.

Enforced Policy

Mean #contacts per person

Median #contacts per person

Mean #contacts per observation

Median #contacts per observation

Overall %contacts

None

0.42

0.16

300.87

115

100.00%

Removing top 2 superspreaders

0.39

0.14

278.56

105

92.58%

Removing top 3 superspreaders

0.38

0.14

269.31

101

89.50

Removing top 4 superspreaders

0.36

0.13

260.94

98

86.72%

Removing top 5 superspreaders

0.35

0.13

253.33

95

84.20%

Removing top 6 superspreaders

0.34

0.13

246.30

92

81.86%

Removing top 8 superspreaders

0.33

0.12

233.59

88

77.63%

Removing top 10 superspreaders

0.31

0.11

222.05

83

73.80%

Removing top 15 superspreaders

0.27

0.10

197.28

73

65.56%

Removing top 20 superspreaders

0.25

0.09

177.04

65

58.84

Removing top 50 superspreaders

0.14

0.04

102.44

33

34.04%

Table 1: Mean and median values of the average number of contacts per person and number of contacts per observation after enforcing the removal of top N superspreaders.
Figure 2: Number of contacts without enforcing any policy and with policies removing the top 2, 10 or 20 superspreaders over time.