面向COVID-19传播模式的多因素影响分析

Multi-factor Influence Analysis for COVID-19 Transmission Patterns

目前,新型冠状病毒肺炎(COVID-19)的传播仍在持续,其传播模式以及影响传播行为的主要因素仍有待深入挖掘。鉴于此,本文从数据分析的角度,通过构造一个特殊的多源数据集(包括COVID-19历史数据、气象数据、人口迁徙数据和空间地理信息数据),以此建立多元Poisson.回归模型(类Poisson回归)来着重分析国内疫情的病毒传播模式及其影响因素。分析结果显示,湿度、平均每日风速、每日的降雨量等气象因素与COVID-19的传播模式显著相关,但与每日温度变化显著不相关。除此之外,COVID-19的传播速度及传播范围与武汉迁出目的地的人口比例、迁入武汉来源地的人口比例以及武汉与其他城市的空间距离均有一定的关联性。全文可视化及模型分析的R代码见:https://github.com/thwgithub/COVID-19 Rcodes.

At present,COVID-19 is still spreading,and its transmission patterns and the main factors for affecting transmission behavior need to be further mined.To this end,from the perspective of data analysis,a multivariate Poisson regression model(Quasi-Poisson)based on an elaborate multi-source dataset(including COVID-19 historical data,meteorological data,population migration and spatial ge-ographic information data),is established,whice focuses on analyzing the transmission patterns and exploring some impact factors for the COVID-19 outbreak in China.The analysis results showed that meteorological factors such as humidity,average daily wind speed,and daily rainfall are significantly correlated with the transmission patterns of COVID-19,but not with the daily temperature change.In addition,the transmission speed and transmission range of COVID-19 are related to the proportion of the population for the destination of Wuhan out migration,the origin of in-migration Wuhun and the spatial distance between Wuhan and other cities.All of the R codes for this paper are available on the site https://github.com/thwgithub/COVID-19-Rcodes.