Results

Comparison of traffic (2019 vs 2020)
Our study analyzed traffic volumes alongside trends in the spread of COVID-19 after the first COVID-19 patient in South Korea was detected. Traffic was analyzed in terms of the number of vehicles operating nationwide. The difference in nationwide traffic between 2019 and 2020 is displayed as ‘Traffic gap (2019 vs 2020)’ in Figure 4, corresponding to the gray shaded area.
Figure 4 Traffic trends based on VDS in 2019 and 2020, and COVID-19 trends in 2020 by region. Data are presented from January 1 to March 31, 2020, on (a) national and (b–f) regional scales. The left y-axis corresponds to traffic and the right y-axis corresponds to the number of confirmed COVID-19 patients. The bold red line corresponds to the traffic trend curve for 2020, and January 19 indicates the first confirmed case in South Korea. The gray dotted line is the difference in traffic between 2019 and 2020, the blue data points are the newly confirmed COVID-19 cases, and the green data points are the cumulative numbers released from isolation.
During the first 3 weeks of 2020, the traffic was around 7% lower than in 2019 (first week −6.7%, second week −0.4%, third week −2.6%). However, following the first confirmed COVID-19 case in South Korea on January 19, 2020, the fourth week of January in 2020 showed a 17.3% increase in nationwide traffic compared with 2019.
In the first week of February, nationwide traffic was 23.3% lower than in 2019. Thereafter, nationwide traffic continued to decrease – in the fourth week of February it was 26.1% lower than in 2019.
In March 2020, nationwide traffic shifted back to an increasing trend from March 7 onwards, as shown by the 2020 traffic trend curve, displayed as a red line in Figure 4. Compared with the same period in 2019, however, traffic was lower throughout March (first week −25.1%, second week −14.6%, third week −13.7%, fourth week −14.0%, fifth week −22.0%).
The mean daily nationwide traffic between January 1 and March 31 was 143 655 563 vehicles, which was 9.7% lower than the same period in 2019 (159 044 566 vehicles) (Table 1 ).
Table 1 Average traffic per day in 2019 and 2020, and COVID-19 trend per day in 2020.
Date Traffic average per day Gapa (%)b Daily new confirmed cases (N) Released from isolation
2019 2020
Jan – 1st week 145 797 502 135 994 670 −9 802 832 (−6.7%) 0 0
Jan – 2nd week 149 049 737 148 389 105 −660 632 (−0.4%) 0 0
Jan – 3rd week 150 897 726 146 908 915 −3 988 811 (−2.6%) 1 0
Jan – 4th week 149 778 529 185 314 734 +25 844 728 (+17.3%) 10 0
Jan – 5th week 147 251 673 150 482 955 +3 231 282 (+2.2%) 7 0
Feb – 1st week 182 825 475 140 144 295 −42 681 180 (−23.3%) 6 2
Feb – 2nd week 162 747 801 165 831 722 +3 083 921 (+1.9%) 4 7
Feb – 3rd week 151 192 280 142 631 273 −8 561 006 (−5.7%) 176 18
Feb – 4th week 170 090 529 125 730 973 −44 359 556 (−26.1%) 2133 27
Mar – 1st week 164 855 643 123 492 052 −41 363 591 (−25.1%) 4430 117
Mar – 2nd week 154 628 156 132 054 132 −22 574 024 (−14.6%) 1319 713
Mar – 3rd week 158 348 967 136 602 840 −21 746 126 (−13.7%) 713 2611
Mar – 4th week 162 656 743 139 886 152 −22 770 591 (−14.0%) 679 4811
Mar – 5th weekc 176 503 164 137 714 570 −38 788 594 (−22.0%) 308 5567
Average 159 044 566 143 655 563 −201 776 965 (−9.7%)
Data: Public data portal, Korea Expressway Corporation point traffic data (date of access: April 1, 2020) (http://data.ex.co.kr/portal/fdwn/view?type=VDS&num=37&requestfrom=dataset#); Korea Centers for Disease Control and Prevention (KCDC), South Korea COVID-19 press release (KCDC, 2020).
a Gap = average traffic per day (2020) − average traffic per day (2019).
b %: (average traffic per day (2020) ÷ average traffic per day (2019)) × 100.
c March 29, 2020 to March 31, 2020 (3 days).
As shown by the regional traffic trend curves in Figure 4, all regions showed a decreasing trend for traffic in February, which shifted to an increasing trend from March. In particular, there was almost no change in traffic in Seoul, while Incheon showed a continuous decrease in traffic from January that shifted to an increasing trend from the end of February. In Gyeonggi, traffic increased in January, showed a slight decrease after the first COVID-19 case, and then switched to an increasing trend again from March 7. In Sejong, the traffic suddenly increased in March. In Daegu, the traffic decreased significantly compared with other regions in February, and shifted to an increasing trend in March; however, overall traffic was still low.

Changes in traffic and COVID-19 trends
In Figure 4, the first COVID-19 patient in South Korea is indicated by a vertical dotted line on January 19. The daily new confirmed cases are displayed as a blue line, and the cumulative number released from isolation is displayed as green squares, corresponding to the right-hand y-axis.
Following the first COVID-19 patient in South Korea, the traffic trend curve (displayed as a red line) decreased continuously until the first week of March, while the frequency of daily new confirmed COVID-19 cases increased during the same period. Thereafter, the national traffic trend curve shifted to an increasing trend from March 7, while the daily new confirmed COVID-19 cases shifted to a decreasing trend, and the cumulative number released from isolation began to show an increasing trend. Thus, as the COVID-19 situation in South Korea began to improve after March 7, the rate of increase in the traffic trend curve continued to grow.
When the regional traffic trend curves and COVID-19 trends were analyzed in Seoul, Incheon, and Gyeonggi, the number of new confirmed cases and the traffic trends in March both increased compared with February. Other regions showed a decrease in the number of new confirmed cases compared with February, while the traffic trends increased (Figure 4).
Scatter plots were created to show regional daily traffic in 2020 against the daily new confirmed COVID-19 cases (Figure 5 ), and single regression analyses were performed (Table 2 ). In Incheon, there was a positive but insignificant linear relationship (β = 43 146; p = 0.056) with an increasing number of new confirmed cases associated with increased traffic. Meanwhile, all the other regions showed negative linear relationships, with traffic decreasing as the numbers of new confirmed cases increased (Figure 5). The regions showing significant linear relationships were the national region (β = −52 176, p < 0.001), Busan (β = −17 895, p < 0.001), Daegu (β = −1778.5, p < 0.001), Gwangju (β = −39 368, p = 0.005), Ulsan (β = −77 689, p = 0.003), Chungbuk (β = −637 223, p = 0.002), Gyeongbuk (β = −49 467, p < 0.001), and Gyeongnam (β = −230 313, p = 0.006) (Table 2).
Figure 5 Scatter plots and single regression lines by region.
Table 2 The result of single linear regression between traffic in 2020 and newly confirmed COVID-19 cases.
β β t-value p
(a) National −52 176.0 −4.17 <0.001***
(b) Seoul −3 025.6 −0.72 0.474
(c) Incheon 43 146.0 1.94 0.056
(d) Gyeonggi −19 180.0 −0.30 0.766
(e) Busan −17 895.0 −3.68 <0.001***
(f) Daegu −1 778.5 −5.58 <0.001***
(g) Gwangju −39 368.0 −2.9 0.005**
(h) Daejeon −71 490.0 −1.66 0.100
(i) Ulsan −77 689.0 −3.03 0.003**
(j) Sejong −806.5 −1.84 0.069
(k) Chungbuk −637 223.0 −3.23 0.002**
(l) Chungnam −62 733.0 −1.96 0.053
(m) Jeonbuk −322 490.0 −1.03 0.308
(n) Jeonnam −217 346.0 −1.15 0.255
(o) Gyeongbuk −49 467.0 −5.05 <0.001***
(p) Gyeongnam −230 313.0 −2.81 0.006**
*p < 0.05, **p < 0.01, ***p < 0.001.

Types of relationship between regional traffic and COVID-19
In Table 3 the analyses in Table 2, Figure 4, and Figure 5 have been classified into types for each region, based on whether the trends in traffic and COVID-19 were increasing or decreasing. Incheon was categorized as a region requiring strong control (Type 1), with increasing trends for both COVID-19 spread and traffic. Gyeonggi and Seoul were categorized as regions in the early stages of focused control or requiring control (Type 2), with increasing traffic but a relatively stable trend for new confirmed COVID-19 cases. The other regions were categorized as stable (Type 3), with increasing traffic but decreasing trends for COVID-19 spread.
Table 3 The level of relationship between traffic and COVID-19 in cities, 2020.
Trend in 2020 Specific City
Level Traffic COVID-19
1 + + (Danger) Strong control required Incheon
2 0 (Caution) Control required, or in the early stage of focused control Gyeonggi, Seoul
3 − (Stable) Under stable control Daegu, Busan, Gwangju, Daejeon, Ulsan, Sejong, Chungbuk, Chungnam, Jeonbuk, Jeonnam, Gyeongbuk, Gyeongnam
+ = increasing; 0 = same; − = decreasing.