Data

Data are available in google drive, as well as Rstudio cloud. In fact this page is built in Rstudio Cloud. You can create an Rstudio account and access the entire workspace so as to write code and copy results like data or graph together.
Link to this project: https://rstudio.cloud/project/3183796

Settings

1. Set a day0
2. time period of interest: T days, say T=30
3. City of interests: the ones that listed in our memo

Initial models

\(Y\)=Cumulative case increase after T days
\(X\)=Number of days with mask mandate during day 0~T-1
\(Y_0\)=Positive/Cumulative cases at day0
Z=some feature of the city(TBD)

\[Y\sim X(+Y_0+Z)\] #### Note

• I am not sure where should we involve the city population. Let \(Y,Y_0\) divided by population?
• Do we need to add more data by using different day0? Of course, more data the better but I think it may have negative impact.
• Data set has test number and positive test rate. I think this is also something important, as test number between different cities may varies
• Reason behind \(Y_0\) is we change assumption to say the transmission mechanism is related to cases. I am afraid there will be . Also, clearly \(Y_0\) is a confounder for X and Y, since higher cases are more likely force the government to anounce stricter mandate
• If we really fancy complex model later on, we may introduce generalized/hierarchical model for different distribution assumption on \(Y\), but that will be another story and does not worth dicussing this week.