Applied Data Science & Machine Learning on Chicago Crimes Dataset
Alejandro Garcia
6/12/2017
INTRODUCTION & MOTIVATION
For the past years, Chicago has been considered one of the most dangerous cities in the US. The motivation behind this project is to help the Chicago Police Department make better decisions with the use of data. Its important for them to receive key insights from this dataset. This dataset contains information from the Chicago Police Department from 2012 to 2017
DESCRIPTION OF DATASET
ID - Unique identifier for the record.
Case Number - The Chicago Police Department RD Number (Records Division Number), which is unique to the incident.
Date - Date when the incident occurred. this is sometimes a best estimate.
Block - The partially redacted address where the incident occurred, placing it on the same block as the actual address.
IUCR - The Illinois Unifrom Crime Reporting code. This is directly linked to the Primary Type and Description. See the list of IUCR codes at https://data.cityofchicago.org/d/c7ck-438e.
Primary Type - The primary description of the IUCR code.
Description - The secondary description of the IUCR code, a subcategory of the primary description.
Location Description - Description of the location where the incident occurred.
Arrest - Indicates whether an arrest was made.
Domestic - Indicates whether the incident was domestic-related as defined by the Illinois Domestic Violence Act.
Beat - Indicates the beat where the incident occurred. A beat is the smallest police geographic area – each beat has a dedicated police beat car. Three to five beats make up a police sector, and three sectors make up a police district. The Chicago Police Department has 22 police districts. See the beats at https://data.cityofchicago.org/d/aerh-rz74.
District - Indicates the police district where the incident occurred. See the districts at https://data.cityofchicago.org/d/fthy-xz3r.
Ward - The ward (City Council district) where the incident occurred. See the wards at https://data.cityofchicago.org/d/sp34-6z76.
Community Area - Indicates the community area where the incident occurred. Chicago has 77 community areas. See the community areas at https://data.cityofchicago.org/d/cauq-8yn6.
FBI Code - Indicates the crime classification as outlined in the FBI’s National Incident-Based Reporting System (NIBRS). See the Chicago Police Department listing of these classifications at http://gis.chicagopolice.org/clearmap_crime_sums/crime_types.html.
X Coordinate - The x coordinate of the location where the incident occurred in State Plane Illinois East NAD 1983 projection. This location is shifted from the actual location for partial redaction but falls on the same block.
Y Coordinate - The y coordinate of the location where the incident occurred in State Plane Illinois East NAD 1983 projection. This location is shifted from the actual location for partial redaction but falls on the same block.
Year - Year the incident occurred.
Updated On - Date and time the record was last updated.
Latitude - The latitude of the location where the incident occurred. This location is shifted from the actual location for partial redaction but falls on the same block.
Longitude - The longitude of the location where the incident occurred. This location is shifted from the actual location for partial redaction but falls on the same block.
Location - The location where the incident occurred in a format that allows for creation of maps and other geographic operations on this data portal. This location is shifted from the actual location for partial redaction but falls on the same block.
DATA WRANGLING
This data came up pretty clean, however, I did apply some simple data wrangling. I eliminated the missing values from this dataset as well as changing the categorical variables to our advantage. This process makes it easier for our analysis later on
crimes_12_to_17_raw <- read.csv("Chicago_Crimes_2012_to_2017.csv", stringsAsFactors = FALSE)
crimes_12_to_17 <- na.omit(crimes_12_to_17_raw)
crimes_12_to_17$Primary.Type <- as.factor(crimes_12_to_17$Primary.Type)
crimes_12_to_17$Description <- as.factor(crimes_12_to_17$Description)
crimes_12_to_17$Location.Description <- as.factor(crimes_12_to_17$Location.Description)
crimes_12_to_17$IUCR <- as.factor(crimes_12_to_17$IUCR)
crimes_12_to_17$Arrest[which(crimes_12_to_17$Arrest == "True")] <- 1
crimes_12_to_17$Arrest[which(crimes_12_to_17$Arrest == "False")] <- 0
crimes_12_to_17$Domestic[which(crimes_12_to_17$Domestic == "True")] <- 1
crimes_12_to_17$Domestic[which(crimes_12_to_17$Domestic == "False")] <- 0
head(crimes_12_to_17)## X ID Case.Number Date Block
## 1 3 10508693 HZ250496 05/03/2016 11:40:00 PM 013XX S SAWYER AVE
## 2 89 10508695 HZ250409 05/03/2016 09:40:00 PM 061XX S DREXEL AVE
## 3 197 10508697 HZ250503 05/03/2016 11:31:00 PM 053XX W CHICAGO AVE
## 4 673 10508698 HZ250424 05/03/2016 10:10:00 PM 049XX W FULTON ST
## 5 911 10508699 HZ250455 05/03/2016 10:00:00 PM 003XX N LOTUS AVE
## 6 1108 10508702 HZ250447 05/03/2016 10:35:00 PM 082XX S MARYLAND AVE
## IUCR Primary.Type Description Location.Description
## 1 0486 BATTERY DOMESTIC BATTERY SIMPLE APARTMENT
## 2 0486 BATTERY DOMESTIC BATTERY SIMPLE RESIDENCE
## 3 0470 PUBLIC PEACE VIOLATION RECKLESS CONDUCT STREET
## 4 0460 BATTERY SIMPLE SIDEWALK
## 5 0820 THEFT $500 AND UNDER RESIDENCE
## 6 041A BATTERY AGGRAVATED: HANDGUN STREET
## Arrest Domestic Beat District Ward Community.Area FBI.Code X.Coordinate
## 1 1 1 1022 10 24 29 08B 1154907
## 2 0 1 313 3 20 42 08B 1183066
## 3 0 0 1524 15 37 25 24 1140789
## 4 0 0 1532 15 28 25 08B 1143223
## 5 0 1 1523 15 28 25 06 1139890
## 6 0 0 631 6 8 44 04B 1183336
## Y.Coordinate Year Updated.On Latitude Longitude
## 1 1893681 2016 05/10/2016 03:56:50 PM 41.86407 -87.70682
## 2 1864330 2016 05/10/2016 03:56:50 PM 41.78292 -87.60436
## 3 1904819 2016 05/10/2016 03:56:50 PM 41.89491 -87.75837
## 4 1901475 2016 05/10/2016 03:56:50 PM 41.88569 -87.74952
## 5 1901675 2016 05/10/2016 03:56:50 PM 41.88630 -87.76175
## 6 1850642 2016 05/10/2016 03:56:50 PM 41.74535 -87.60380
## Location
## 1 (41.864073157, -87.706818608)
## 2 (41.782921527, -87.60436317)
## 3 (41.894908283, -87.758371958)
## 4 (41.885686845, -87.749515983)
## 5 (41.886297242, -87.761750709)
## 6 (41.745354023, -87.603798903)ANALYSIS ON DATASET
#install.packages('ggplot2')
library(ggplot2)
primary_type <- ggplot(crimes_12_to_17, aes(Primary.Type))
primary_type + geom_histogram(stat = "count") + coord_flip()## Warning: Ignoring unknown parameters: binwidth, bins, pad
It is clear from this graph that the most common types of crimes are - Theft - Battery robbery - Criminal damage - Narcotics
ggplot(crimes_12_to_17, aes(Year)) +
geom_density()
From this density plot we can observe how crimes have decreased since year 2012, however it has not decreased drastically from 2014-2016
#install.packages('plotrix')
library(plotrix)
arrests <- table(crimes_12_to_17$Arrest)
lbls <- paste(names(arrests), "\n", arrests, sep="")
pie3D(arrests, labels = lbls,
main="Arrests results (1 = True, 0 = False) from Crimes commited ")
Out of all the crimes committed, only 26% of the crimes resulted in an arrest
domestic <- table(crimes_12_to_17$Domestic)
lbls <- paste(names(domestic), "\n", domestic, sep="")
pie(domestic, labels = lbls,
main="Domestic results (1 = True, 0 = False) for Crimes commited ")
Out of all the crimes, only about 15% where domestic
#levels(crimes_12_to_17$IUCR) #353 Levels
top10_iucr <- tail(names(sort(table(crimes_12_to_17$IUCR))), 10)
iucr_raw <- table(crimes_12_to_17$IUCR)
barplot(iucr_raw[order(iucr_raw, decreasing = TRUE)], xlim = c(0,11))
The most common IUCR Codes - Theft of $500 and under - Domestic battery - Battery robbery - Damage to property
#levels(crimes_12_to_17$Description) #340 Levels
top10_description <- tail(names(sort(table(crimes_12_to_17$Description))), 10)
head(top10_description)## [1] "FROM BUILDING" "AUTOMOBILE"
## [3] "FORCIBLE ENTRY" "POSS: CANNABIS 30GMS OR LESS"
## [5] "TO PROPERTY" "OVER $500"It is clear that the top Descriptions of crimes are: Building, automobile, and forcible entry. This is key for the Chicago Police Department
#levels(crimes_12_to_17$Location.Description) #141 Levels
top10_location_description <- tail(names(sort(table(crimes_12_to_17$Location.Description))), 10)
head(top10_location_description)## [1] "SCHOOL, PUBLIC, BUILDING" "SMALL RETAIL STORE"
## [3] "RESIDENTIAL YARD (FRONT/BACK)" "ALLEY"
## [5] "PARKING LOT/GARAGE(NON.RESID.)" "OTHER"#street, residence, apartment, etc...
#location_description_raw <- table(crimes_12_to_17$Location.Description)
#barplot(location_description_raw[order(location_description_raw, decreasing = TRUE)], xlim = c(0,11))
#scaling the graph is impossibleThe top locations where crimes occur include school, retail stores, and residentail yards. This is important because the police can pay more attention to these areas
crimes_12_to_17$Beat <- as.factor(crimes_12_to_17$Beat) #Put this at the beggining of the report
#levels(crimes_12_to_17$Beat) #289 Levels
top10_beat <- tail(names(sort(table(crimes_12_to_17$Beat))), 10)
beat_raw <- table(crimes_12_to_17$Beat)
barplot(beat_raw[order(beat_raw, decreasing = TRUE)], xlim = c(0,11))
Most common beat (a small police geographical area). Additional reference for beats: https://data.cityofchicago.org/Public-Safety/Boundaries-Police-Beats-current-/aerh-rz74
crimes_12_to_17$District <- as.factor(crimes_12_to_17$District) #Put this at the beggining of the report
#levels(crimes_12_to_17$District) #23 Levels
top10_district <- tail(names(sort(table(crimes_12_to_17$District))), 10)
district_raw <- table(crimes_12_to_17$District)
barplot(district_raw[order(district_raw, decreasing = TRUE)], xlim = c(0,11))
These are the most common Districts where crimes occur in Chicago - District 11 - District 8 - District 6
crimes_12_to_17$Ward <- as.factor(crimes_12_to_17$Ward) #Put this at the beggining of the report
#levels(crimes_12_to_17$Ward) #45 Levels
top10_ward <- tail(names(sort(table(crimes_12_to_17$Ward))), 10)
ward_raw <- table(crimes_12_to_17$Ward)
barplot(ward_raw[order(ward_raw, decreasing = TRUE)], xlim = c(0,11))
The Ward is simply the City Council District where the crimes occurs. Additional visual reference: https://data.cityofchicago.org/d/sp34-6z76.
crimes_12_to_17$Community.Area <- as.factor(crimes_12_to_17$Community.Area) #Put this at the beggining of the report
#levels(crimes_12_to_17$Community.Area) #67 Levels
top10_community_area <- tail(names(sort(table(crimes_12_to_17$Community.Area))), 10)
community_area_raw <- table(crimes_12_to_17$Community.Area)
barplot(community_area_raw[order(community_area_raw, decreasing = TRUE)], xlim = c(0,11))
These areas also represent where the most common crimes occur. As you might observe, Area 25 is increasingly high compared to other areas. This is a great piece of information that the police should now and pay attention to.
crimes_12_to_17$FBI.Code <- as.factor(crimes_12_to_17$FBI.Code) #Put this at the beggining of the report
#levels(crimes_12_to_17$FBI.Code) #19 Levels
top10_fbi_code <- tail(names(sort(table(crimes_12_to_17$FBI.Code))), 10)
fbi_raw <- table(crimes_12_to_17$FBI.Code)
barplot(fbi_raw[order(fbi_raw, decreasing = TRUE)], xlim = c(0,11))
The most common FBI Codes - Larceny - Simple Battery - Vandalism
RESULTS AND RECOMMENDATIONS
From the previous graphs we were able to obtain significant information like: - Most common type of crimes - Decrease in crime rate from year 2012-2017 - Crimes that resulted in Arrests - Top IUCR Codes - Most common areas where crimes occur - Top FBI Codes
Based on the analysis done on the Chicago crime scene data set, we are able to offer several recommendations to the Chicago Police Department Focus on the most common types of crimes in order to prevent them More arrests have to be made per crime, only 26% of the criminals where arrested Theft of $500 in the streets is very common, definitely have more patrols going around the area Set a specific perimeter for the areas where crime scene is extremely high