We started with scrapping reviews of 300 closed restaurants in San Francisco, CA in order of their appearance on google search. The first task was to find out, of the collected restaurants which restaurants actually had any reviews to be a vaild data point for our research.
We were able to find out 92.3% of restaurants (277) did have reviews so our sample data was good enough to work with. Also we took into account if a restaurant had way to many reviews our data processing time was increased to much. To take care of that we limited our reviews to 60 per restaurant for our anlysis.