Chapter 5 Results

We have deeply understand that correlation does not represent causal relationship. Actually, due to the existence of confounding bias and selection bias, overal data distribution can even direct to the totally opposite conclusion. We are interested in exploring the correlation behind the data and have tried to interpret them by drawing attractive and insightful graph, and we pay much attention on extracting possible causal relationship to get a better understanding and present potential suggestions to avoid crimes, and thus create a safer and relaxing community.

5.1 Crime number vs category of crimes among different specific locations.

The incidences of harrassment are outstandingly high among all the types of crimes, with that of assult and petite larceny being the next highest. All of these three types of crimes are not severe and are still being a big part of social problems now in 2021. Another group that is outstanding is the 25-44 age group, with the largest number of victims among all the age groups. In addition, it can also be easily seen that Staten Island has the lowest incidents among all types of crimes.

5.2 Trend of number of crimes over years

While no remarkable contrast between the number of crimes among the three years, some minor differences can still be perceived. In 2020, the number of crimes slightly decreases compared to the former two years, especially for larceny, harrassment and assult. This has shown that various orgnizations, institutions, and the government have indeed taken effective actions aiming to solve these issues and this graph has shown that there has been positive progresses.

5.3 Parallel coordinates chart on crime types and locations

The above parallel coordinates chart displays the number of crime incidences of the top 20 crime types, grouped by specific locations each crime incidence took place and the borough where each incident occurred.

It can be seen that among the four boroughs of New York City, Staten Island has the least incidences of crimes in general, with the most occurred type of incidents being intoxicated and impaired driving. Queens also has relatively low crime incidences. The most occurred incidences are grand larceny of motor vehicle. Starting from Manhattan, the number of crime incidents starts to increase. Apartment houses are the No.1 places for crimes such as theft-fraud and grand larceny to happen. Grand larceny also takes place in Manhattan streets. Brooklyn and Bronx both have very high crime rates and almost peak out at the number of almost evrey type of crimes, wth most of them taking places in the streets and apartment houses.

For most of the specific locations that a crime can take place at, the number of each type of crime is consistent with other types for most of the cases, that is, a specific location has either a generally very low (or moderate) crime rate or a generally high crime rate. For instacne, public schools, hospitals, hotels, and fast food shops all have overall very low crime rate; on the other hand, places like streets and apartment houses have the highest overall crime rate. However, there is one exception: chain stores have relative small number of crimes under all the crime categories but very high instances of petite larceny.

5.4 Visualization about portraits of suspects and victims.

We want to see the portrait of suspects and victims of all crimes in recent three years. By an alluvium, we focus on the victim sex, denoted by different colors, as well as the other information related to one crime event. From the graph above, we can have a direct understanding that crimes in NYC happen on women more than men and as for the suspect number, the number of male suspect is 3 times as women. In addition, it is not hard to see that in male and female suspects, their targets both focus more on women. Certainly there is no gender opposition in crimes, but we do see women is always the vulnerable group.

STOP WOMEN VIOLENCE! More help and guard shoule be paid on women.

We have understood that there is a specific age structure in the whole population and we do not gather related information, but we do see that in NYC, compared between victim and suspects, old people aged from 45 to 64 is more vulnerable.

5.5 Visualization about daily crime trend.

In this graph, we see the daily trend of crime number in five counties. The y axis is the crime number, the x axis is the hour in a day. The fill part is the victim sex. Since the population and areas of five counties are unknown, we free the y axis and only focus on the daily trend. And we do find a very similar trend in these five counties. The daily valley appears at 5-7 o’clock while the daily peak appeas at 15-19 o’clock. (We exclude the 12 o’clock peak since we regard it as the auto fill of missing value and it does not meet the daily trend) Besides, we add victim sex in the graph to show more solicitude for sex discrepancy. Unfortunately, we see that in all these five counties, in daily time, women are more vulnerable than men. However, at night, from appr. 22 o’clock to 5 o’clock, men are as much vulnerable as women. In staten island this conclusion is not obvious while in other four counties, the crime numbers of men and women are nearly the same.