Conclusions
After conducting exploratory analysis, clustering analysis, time series, and OLS regression. I found that the restaurant hygiene grading system is averaged the percentages of each grade in every boroughs. But within each boroughs, there are differences that might correlate with population density and median incomes. Besides, the time series show that the number of inspections has been increased in the past four years while the number of restaurants remains relatively unchanged. It means that the inspectors from the department have conducted inspections more frequently. The time series of monthly trend also indicates that during Summer and Winter seasons, the number of inspections is decreased a lot. This might result from the dramatic change of temperature and the holiday season as well. At the end, the OLS regression reveals that the time of inspections has little impact on the scores, while the population and cluster labels have more influence on evaluating the score.
The weakness of the project is that it omits other important factors such as median incomes, food accessibility, and environments. Further works on these topics can improve this analysis.