Fig 1.
Infection cases in Tokyo MA (A) and entire Japan (B).
Fig 2.
Outline of mobile spatial statistics data.
Fig 3.
Comparison of temporary population at 10:00 a.m. to 9:00 and 11:00 a.m.
The data includes the average population on Thursdays in November 2019 in Tokyo MA.
Fig 4.
Complementary cumulative distribution function of population in approximately 500m × 500m grid cells.
Fig 5.
Location of high-density places.
(A) Black dots in the map represent grid cells in which the average population at 10:00 a.m. on Thursdays in November 2019 exceeded 4,000. The yellow area represents Tokyo MA. (B) We analyze four prefectures: The deep green is Kanagawa prefecture, the yellow is Tokyo prefecture, The orange is Saitama prefecture, and the gray is Chiba prefecture. We obtained Figs 5A and 5B from Natural Earth and National Land Information Division, National Spatial Planning and Regional Policy Bureau, Ministry of Land, Infrastructure, Transport and Tourism of Japan.
Fig 6.
Population density in Tokyo MA in comparison to before COVID-19: (A)Before COVID-19; (B)First lockdown; (C) Second lockdown.
We obtained Figs 6A, 6B, and 6C from Natural Earth and National Land Information Division, National Spatial Planning and Regional Policy Bureau, Ministry of Land, Infrastructure, Transport and Tourism of Japan.
Table 1.
Baseline results.
Fig 7.
Total marginal effects of infection cases (IFC).
Population density and trip distance are fixed to 2000 and 10km when the other variable is changed.
Fig 8.
Total marginal effects of the lockdowns: (A)First lockdown; (B) Second lockdown.
The total marginal effects of each lockdown are described by β0m + β1mDist + β2mDen, where β0m, β1m, and β2m represent the coefficients of the single and cross terms of the mth lockdown in Model 2 of Table 1. The dashed lines show the total effects of the IFC under the same conditions. Those with the first lockdown are multiplied by 20 in the left, Population density, and 30 in the right, Trip distance (m). However, no scale adjustment is performed for the second lockdown.
Fig 9.
Date fixed effect: (A) in Model 3 and (B) in Model 4.
In both figures, date fixed effect significantly decreases around April, 2020.
Table 2.
Estimation results for detailed periods.
Fig 10.
Total marginal effects divided into five periods.
We also calculate total marginal effect as for IFC, LD1, and LD2 by Eq 3.
Table 3.
Estimation results for the weekend afternoon.