New York Crime Number Changing
Tracing from 2001 —— 2015, What happened on 2013 Oct 31st?
October 12, 2017
LIS-658-01 Tableau Public Lab
Lab two, mainly to learn the Tableau Public and use the Tableau to build some visualization based on the data and the topic we chose. My topic is about the crime report recording in New York City from 2001 to 2015. I got the data from Data.gov(https://catalog.data.gov). Hope to know the tendency of crime happened in New York, what time during a day is dangerous, if the daytime is really safer than the night time? And perhaps, according to the data, we could find some weird milestone of the crime reported. As for the reason to choose the crime data of New York City, is because I live in New York and New York somehow as a focus point, especially its crime is often the worth to discuss.
Three Inspired Visualization
In fact, Fig.0 here only show a part of the first example of the visualization I found. Fig.0 is only the bottom part of a poster, a visualization talking about the Beatles. Because it is a long poster, so I only cut this part, which inspired me a lot. Due to my purpose is to show the crime report number changing during the date, and I also hope that my viewer could even see how it changes during the hours of a day. Therefore, I decided to use two axes to describe the time and the visualization examples here helps me come up the idea, that I could use color and shapes to represent more information in my diagram.
Then, I got more detailed idea from Fig.2. According to the data I found, the crime has been divided into 3 category, felony, misdemeanor and violation. Except to use different color to separate them, I also want to put them together to directly see the comparison between the three kinds of crime at the same time. Then the pie chart shows in the in the Fig.1 works well with the idea.
As for Fig.3, because I also have the data about where the crime happens, so I’m also considering making some map visualization as Fig.2 showing to see how the crime distribute in New York City on map. But due to the time I did not make this idea come true yet.
Materials and Methods
My data is the New York Crime data recoded from 2001 — 2015. It is an csv file, you could find it by searching with the key word New York Crime, csv. And you could find the dataset named Crimes – 2001 to present on the Data.gov website. This csv data file contains totally 22 columns. From the left to the right, they are the number for each order, named Cmplnt_Num, then five different columns introduce five different time, include report date, time, and so on. Then is mainly about the type of the crime, include the describe to the crime, one column with compactly words, like Forgery, Assault 3 and so on. 4 columns introduce this part, then the next one record if the crime succeeded or not. After this, before the location part, the file use two columns describe the level of the crime, like felony, misdemeanor and violation. The other one tells which organization handle on this report issues, like N.Y. POLICE DEPT. At last, the rest 11 columns are related to the geo information. Include the district where it happens, like Queen, Brooklyn and so on; Address number, inside or outside of the building, what the building use for, a bar, a restaurant and so on; Longitude, latitude, and so on. Generally, this crime data file has a very detailed information for those crimes recorded in it. And it is possible to use this data to trace the crime reported number change during the time, and even draw some crime data map based on these data.
What’s more, the software we used for the lab this time is named Tableau, a professional data visualization tool which can help people build interactive data visualization. In addition, their product also focused on business intelligence. We choose it public free service to build our visualization.
As the result, the visualization I build through the Tableau Public is as followings. I choose three colors to represent the different kind of the crime. The red orange means the felony, yellow orange means the misdemeanor and the light green means the violation. I choose the red here because red often use to represent the serious meaning, give some orange for the better visual feelings, and the yellow orange hope to give the sense that this is a kind of crime but not so bad as the felony. The light green hope to show the totally opposite of the felony or the misdemeanor. Because this is most slightly crime there, I hope to compare it to the other real crime there. The color key shows as the Fig.3.
And from the Fig.4, we could see that from 2001 to 2013, crime happens very seldom, and seems focus happened around the midnight, 0am. I think this also means that for most people, they will be safe in New York except the midnight during 2001 to 2013. And in fact, even at that time midnight happens a lot of crime but it cannot match what happens in New York after 2013. We could see that after 2013, the number of crime reported increase sharply. And not only in the midnight, midnight still will be the most possibility time to happen the crime, but the other time are also looks become very dangerous. (Fig.5 shows a detailed visual of the data from 2001 – 2015, and focus on the time from 0am – 6am).
In addition, as the Fig.6 showing, I try another visualization to show the huge change of the number of the crime reported. And from this image, we could also see that the crime of Misdemeanor occupied the most percentage of the number of the report, the violation only occupied a very limited part. And for example, the number of those report happens in 2012 is 678 reports, and only have the felony, but in 2013, the number changed to 28792 felonies, 47432 misdemeanors, and 9938 violations, then for 2014, the number increased to 150716 felonies, 276770 misdemeanors and 62877 violations.
Then I look closely to the data at the end of the 2013, and just as the Fig.7 showing, between the 10/31/2013 to the 11/01/2013, the number changed intensely. Looks like something, some milestone happened on that day. And I only found this kind of phenomenon on this day, after this, even the total number of the crime increase sharply, but the data looks like have a trend of gradually rising. The data do not change immediately on any special day. Then the Fig.8 also could prove this, Fig.8 show the crime report data recorded from 2013 – 2015.
What’s worse, out of curiosity, I made the Fig.9 to check the status for all the reported crime from 2001 – 2015. This visualization shows a disheartening information that for all three kinds of the crime reported in the data, when it has been reported, the crime has already completed, only very limited part has been signed as the attempted. The number are as followings: 327678 felonies reported but only 13224 reported when it is still attempted, the rest part all reported after it completed. 585384 misdemeanors and only 5184 attempted. 134937 violations, 0 attempted.
Another impressive thing I found through the visualization is that it is very hard to find the violation happens in the data, it just takes very limited part through the whole record. In my opinion, it shows that for the crime, the different level does not have progressive relationship. Seems like people will not against the violation, then try misdemeanor and at last felony. I’m wondering maybe it is because the crime is most related to the motivation to do something, or violation is hard or mostly do not need to be reported.
This time is a good experience of using the tableau to build the data visualization. For the visualization part, I think it is better to go to the direction to build the geo inform part. Draw the map to show where is dangerous, where happens more crime. And it is inside, outside, in a bar or on the street. With those information, maybe this study will be more useful for viewers to avoid dangerous. And for the history part, maybe I could start a research on what happens on 10/31/2013 and then lead the increase of crime. And maybe could also try to find the reason why more illegal actions than the violation.