Trends in Seattle Crimes¶
Danielle Shen¶
Intro to Project:¶
Seattle is an increasingly growing city due to the boom in tech giant headquarters built there. As a result of these companies, the rise in the cost of living keeps increasing, which could lead to higher crime rates as the city becomes increasingly gentrified. This project will analyze data from the Seattle Police Department's public data sets from 2008 to 2020.¶
In [1]:
from IPython.display import Image
In [2]:
Image("seattle-travel.jpg")
Out[2]:
Seattle's famous Space Needle¶
Many giant tech corporations have offices in Seattle, including Facebook, Microsoft, Google and Amazon to name a few. But with the influx of employees finding living in the city, the rise of living increases as the city takes advantage of the new influx of wealthy transplants by building new apartment complexes and raising rent in existing buildings. This gentrification has devastating impacts on the local community, who can't afford to live in the area anymore and are forcedd out.¶
To understand more about what native population is fighting for, refer to this recent publication: https://www.newsweek.com/seattle-capitol-hill-protesters-gentrification-black-lives-matter-1511192¶
The crime rates in a gentrified city also change like the population. Theorists have conflicting arguments on whether or not gentrification causes crime rates to increase or decrease. It is my hypothesis that crime rates have increased in Seattle due to gentrification and I will be using data anlysis to prove my theory.¶
Python Libraries used:¶
- Folium
- Requests
- Pandas
- Numpy
- datetime
- matplotlib
To get a better understanding of the visualization tool folium, this document provides in-depth explanations of it's basic functions:
To get a better understanding of the pandas library, this guide listed below is a quick and effiecient cheat sheet:
The CSV file used containing all the data in this project is from Seattle's official city government site. The site contains citizen, business, and visitor information sections, plus city government information. Specfically, the section used is it's Public Data Sets. In that section, it contains high value, machine-readable datasets created by the City of Seattle available for public use. The file we are using is it's "Crime data set from 2008 to Present." The following process was used to create tidy data. (In tidy data each variable forms a column and each observation forms a row.
In [3]:
!pip install folium
import folium
import matplotlib.pyplot as plt
from sklearn.linear_model import LinearRegression
from sklearn import linear_model
from datetime import datetime
import requests #get request
import pandas as pd #pandas
import numpy as np #
import sqlite3
import matplotlib.pyplot as plt
import seaborn as sns
Retrieving the 2008-Present Seattle Crime Data Set:
In [4]:
url ="https://data.seattle.gov/api/views/tazs-3rd5/rows.csv?accessType=DOWNLOAD"
seattleCrime = pd.read_csv(url)
In [5]:
seattleCrime
Out[5]:
In [6]:
(seattleCrime['Crime Against Category']).unique()
Out[6]:
From the entire crime data set, there are 4 categories the crimes committed against: Society, property, persons or not a crime.¶
In [7]:
(seattleCrime['Offense Parent Group']).unique()
Out[7]:
From the entire crime data set, there are 34 parent groups all the crimes committed fall under.¶
Let's create a scatter plot of the year 2020 mapping the offense parent groups to the crimes against category. This will provide an overall relationship between the two groups:¶
The matplotlib.pyplot.scatter function will be used. For a condensed run down of the functions of matplotlib, their website breaks down the basic functions:¶
1. https://matplotlib.org/3.3.3/api/_as_gen/matplotlib.pyplot.scatter.html¶
In [8]:
array_1=[]
array_2=[]
for x in range(0,3000):
array_1.append(seattleCrime.loc[x]['Offense Parent Group'])
array_2.append(seattleCrime.loc[x]['Crime Against Category'])
plt.figure(figsize=(25,10))
plt.ylabel("Offense Parent Group")
plt.xlabel("Crime Against Category")
plt.title("Offense Parent Group vs Crime Against Category")
plt.scatter(array_2,array_1)
Out[8]:
From this plot above, you can see for the year 2020, the most crimes committed were crimes against property because the dots are the most concentrated for that column.¶
To begin tidying the data, a copy of the crimes table is made to edit and modify:¶
In [9]:
seattleTidy = pd.read_csv(url)
seattleTidy.dropna(inplace=True)
# drop all the rows with Nan
In [10]:
seattleTidy=seattleTidy.drop("Offense End DateTime",axis=1)
seattleTidy=seattleTidy.drop("Offense Code",axis=1)
# drop the dates and times the crimes ended because many are unknown. Drop the offense code because that is not going to be analyzed.
In [11]:
crimeCounts = sorted(seattleTidy["Offense Parent Group"].unique())
# We are going to store the most populous Offense Parent Groups, AKA the most popular groups crimes commited fall under.
# First pull all unique Offense Parent Groups
crimeCounts2=[]
temporary_Seattledf = seattleTidy.copy()
# create a copy of data to iterate through
index = 0
# Function iterates through every element in the seattle crimes dataset to capture its Offense Parent Group
for ele in crimeCounts:
crimeCounts2.append((ele, temporary_Seattledf.groupby("Offense Parent Group").count()["Offense"][0:].values[index]))
index += 1
crimeCounts2 = sorted(crimeCounts2, key=lambda x: x[1], reverse = True)
crimeCounts2[0:10]
# Now only pull the top 10 Offense Parent Groups
Out[11]:
In [12]:
SeattleTop_5_crimes = crimeCounts2[0:5]
SeattleTop_5_crimes
# Save the top 5 most occuring offense parent groups
Out[12]:
In [13]:
# Now to begin looking at the times and dates each crime occured. Because the original data set already
# has the date and crime formatted in YEAR-DAY-MONTH TIME, we only need to capture each part in a regex
# so that the elements for the Offense's Start Date Time are easily accessible
time_format = "%Y-%m-%d %H:%M:%S"
# to save the properties of column 'Offense Start DateTime'
# iterates through entire crime data set to save the elements of 'Offense Start DateTime' for easy access later
for idx in seattleTidy.index:
piInString=str(seattleTidy.at[idx,"Offense Start DateTime"])
seattleTidy.at[idx,"Offense Start DateTime"] = datetime.strptime(piInString,time_format)
seattleTidy.at[idx,"Report DateTime"] = datetime.strptime(seattleTidy.at[idx,"Report DateTime"],time_format)
seattleTidy.head()
Out[13]:
In [14]:
# Now that we have the time each crime occured easily saved in the regex, we can plot the times each crime occured
# for Seattle's top 5 Offense Parent Groups. This function will be able to show us if most crimes happen during
# the day or during the afternoon or night.
Hourlycrimes = {}
for i in range(len(SeattleTop_5_crimes)):
Hourlycrimes[SeattleTop_5_crimes[i][0]] = []
for idx in seattleTidy.index:
t = seattleTidy.at[idx,"Offense Parent Group"]
if t in Hourlycrimes:
hour = seattleTidy.at[idx, "Offense Start DateTime"].hour
Hourlycrimes[t].append(hour)
for i in Hourlycrimes:
hist, bin_edges = np.histogram(Hourlycrimes[i], bins=24)
plt.hist(Hourlycrimes[i],bins=24,color="red",edgecolor='black', linewidth=1.2)
plt.ylabel("Number of " + i)
plt.xlabel("Hour of Day")
plt.title(i + " Crimes per Hour")
plt.show()
From all 5 histograms of the 5 most common offense parent groups, from 12 AM to 10 AM is when the least amount of crimes is committed and from 10 AM to 8 PM is when the most crime is committed.¶
We should also map the seasons each crimes occurs because the general preseumption is that crime increases during the winter (December-February) because of higher seasonal depression leading to higher crime rates:¶
In [15]:
seattleTidy['SEASON OF OCCURENCE'] = None
# initialize the season that the crime occured column to nothing
# Iterate through the data table and we can easily pull from our regex what month the crime occured in
# from the data column "Offense Start DateTime"
for idx in seattleTidy.index:
curr_month = seattleTidy.at[idx, "Offense Start DateTime"].month
# grabs the month
if curr_month == 12 or (1 <= curr_month and curr_month <= 2):
seattleTidy.at[idx,'SEASON OF OCCURENCE'] = "Winter"
# months 12, 1 and 2 are Winter
elif 3 <= curr_month and curr_month <= 5:
seattleTidy.at[idx,'SEASON OF OCCURENCE'] = "Spring"
# months 3, 4, and 5 are Spring
elif 6 <= curr_month and curr_month <= 8:
seattleTidy.at[idx,'SEASON OF OCCURENCE'] = "Summer"
# months 6, 7, and 8 are Summer
elif 9 <= curr_month and curr_month <= 11:
seattleTidy.at[idx,'SEASON OF OCCURENCE'] = "Fall"
# months 9, 10 and 11 are Fall
seattleTidy.head()
Out[15]:
In [16]:
seattleTidy['Year Start'] = None
for idx in seattleTidy.index:
curr_year = seattleTidy.at[idx, "Offense Start DateTime"].year
seattleTidy.at[idx,'Year Start'] = curr_year
In [17]:
seattleTidy.head()
Out[17]:
Now we can also map the years the crimes occured in and if crimes are increasing with time. Gentrification in Seattle was not as high as it was in the present year (2020) and crimes rates could have increased due to the rapidly growing cost to live in the changing city:¶
In [18]:
Yearlycrimes2 = {}
# new data structure to map years to crimes committed
for i in range(len(SeattleTop_5_crimes)):
Yearlycrimes2[SeattleTop_5_crimes[i][0]] = []
# iterate through
for idx in seattleTidy.index:
t = seattleTidy.at[idx,"Offense Parent Group"]
if t in Yearlycrimes2:
year = seattleTidy.at[idx, "Offense Start DateTime"].year
Yearlycrimes2[t].append(year)
for i in Yearlycrimes2:
hist, bin_edges = np.histogram(Yearlycrimes2[i], bins=24)
plt.hist(Yearlycrimes2[i],bins=24,color="green",edgecolor='black', linewidth=1.2)
plt.ylabel("Number of " + i)
plt.xlabel("Year")
plt.title(i + " Crimes per Year")
plt.show()
From the histograms comparing the years to the number of crimes committed for the top 5 most common crime parent groups, the year 2020 had the highest occurances of all 5 groups. It can imply that skyrocketing cost of living due to gentrification increases the wealth divide in the the city and crimes occur more frequently.¶
In [19]:
seattleTidy
Out[19]:
To plot comparing the amount of crimes committed each season, a bar graph is a good visualization tool:¶
In [20]:
# This function iterates through the data set to capture the season each crime occured in and adds it to the counter
# variable for each of the 4 seasons.
crimeSeason = sorted(seattleTidy["SEASON OF OCCURENCE"].unique())
crimeSeason2 = []
tempSeason_df = seattleTidy.copy()
# make a copy of data set to iterate through
index = 0
for ele in crimeSeason:
crimeSeason2.append((ele, tempSeason_df.groupby("SEASON OF OCCURENCE").count()["Offense"][0:].values[index]))
index += 1
crimeSeason2
Out[20]:
In [21]:
lis=crimeSeason2
lis[0], lis[-1] = lis[-1], lis[0]
lis
# The list is re-ordered so that the seasons appear in the correct order correct order by swapping the first and last element of the list.
Out[21]:
In [22]:
# Now we need to edit the elements of the list with crimes and seasons mapped so that it can be visualized.
# For each index of the list:
# The tuple with (season,number of crimes) must be turned into a type (string, int) tuple
crimeSeason2Copy=lis
graphOfSeasons = [(elem1, elem2) for elem1, elem2 in crimeSeason2Copy]
for (a,b) in graphOfSeasons:
b=b.item()
print(type(a),type(b))
In [23]:
# Now we can manually create a bar graph and also plot a line graph for seasons vs. crimes committed.
# Here is the code that uses plt functions to plot the bar graph:
fig = plt.figure()
ax = fig.add_axes([0,0,1,1])
langs = ['Winter', 'Spring', 'Summer', 'Fall']
students = [106899,117248,114099,116887]
ax.bar(langs,students)
plt.title("Crimes Committed Per Season From 2008 to 2020")
plt.xlabel("Season")
plt.ylabel("Number of Crimes Committed")
plt.show()
In [24]:
# Must zip list of (season,crimes) to be able to plot using line graph
plt.figure(figsize=(20,10))
plt.plot(*zip(*crimeSeason2Copy))
plt.title("Crimes Committed Per Season From 2008 to 2020")
plt.xlabel("Season")
plt.ylabel("Number of Crimes Committed")
plt.show()
The same process is repeated to plot the 5 most common Parent Offense Groups the most populous crimes committed fall under:¶
In [25]:
crime_copy=SeattleTop_5_crimes
In [26]:
testList2 = [(elem1, elem2) for elem1, elem2 in crime_copy]
for (t,w) in testList2:
w=w.item()
print(type(t),type(w))
In [27]:
fig2 = plt.figure(figsize=(20,10))
# adjust image size
ax2 = fig2.add_axes([0,0,1,1])
TypeParentOffenseGroup = ['LARCENY-THEFT', 'BURGLARY/BREAKING & ENTERING', 'MOTOR VEHICLE THEFT', 'DESTRUCTION/DAMAGE/VANDALISM OF PROPERTY','FRAUD OFFENSES']
# 5 most popular parent offense groups
numC = [220201,62254,45195,42639,35439]
# corresponding number of crimes committed for each group
ax2.bar(TypeParentOffenseGroup,numC)
plt.title("Crimes Committed for Top 5 Offense Parent Groups")
plt.xlabel("Offense Parent Group")
plt.ylabel("Number of Crimes Committed")
plt.show()
# lables
In [28]:
plt.figure(figsize=(20,10))
plt.plot(*zip(*testList2))
plt.title("Crimes Committed in Each Parent Offense Group")
plt.xlabel("Offense Parent Group")
plt.ylabel("Number of Crimes Committed")
plt.show()
From the graphs, we can see that Larceny-Theft is the parent group most crimes fall under, followed by burglary, then motor vehicle theft, destruction of property and lastly fraud offenses. Theft is very reasonable to be the most commonly occuring crime because in an area with high living cost, criminals know to target individuals there because they will likely have disposable income.¶
To look at specific offenses (crimes) committed, from the Offense column analysis, there is a large variety, including Burglary/Breaking & Entering, Destruction/Damage/Vandalism of Property, All Other Larceny, Robbery, Pornography/Obscene Material, Aggravated Assault, Drug/Narcotic Violations, Wire Fraud, Credit Card/Automated Teller Machine Fraud and Driving Under the Influence as the top 10 most occuring.¶
From the analysis above, it makes sense the specific crimes committed are mostly theft crimes, because Larceny was the most common parent group all the crimes fell under. To plot the 5 most common offenses (crimes):¶
In [29]:
a=seattleTidy['Offense']
a
Out[29]:
In [30]:
crimewCounts = sorted(seattleTidy["Offense"].unique())
crimewCounts2=[]
tempw_df = seattleCrime.copy()
index = 0
for ele in crimewCounts:
crimewCounts2.append((ele, tempw_df.groupby("Offense").count()["Offense Parent Group"][0:].values[index]))
index += 1
crimewCounts2 = sorted(crimewCounts2, key=lambda x: x[1], reverse = True)
crimewCounts2[0:10]
Out[30]:
In [31]:
top5Offenses=crimewCounts2[0:5]
In [32]:
top5Offenses
Out[32]:
In [33]:
crimes5=top5Offenses
In [34]:
plt.figure(figsize=(20,10))
plt.plot(*zip(*crimes5))
plt.title("Most Common Offenses")
plt.xlabel("Offense")
plt.ylabel("Number Committed")
plt.show()
Finally we can look at the areas where crimes most commonly occur. Using folium to display the areas where the crimes occured is a great visualization tool.¶
In [35]:
heatMapCSV=seattleTidy
heatMapCSV.dropna(inplace=True)
heatMapCSV
# Create a copy of data set to use for folium maps
Out[35]:
In [36]:
# Because we only care about the crime committed and the longitude and latitude of each crime, many of the prior
# crime information can be dropped to make iterating easier
heatMapCSV=heatMapCSV.drop("Group A B",axis=1)
heatMapCSV=heatMapCSV.drop("Precinct",axis=1)
heatMapCSV=heatMapCSV.drop("Sector",axis=1)
heatMapCSV=heatMapCSV.drop("Beat",axis=1)
heatMapCSV=heatMapCSV.drop("MCPP",axis=1)
heatMapCSV=heatMapCSV.drop("Report DateTime",axis=1)
heatMapCSV=heatMapCSV.drop("Offense ID",axis=1)
heatMapCSV=heatMapCSV.drop("100 Block Address",axis=1)
heatMapCSV=heatMapCSV.drop("SEASON OF OCCURENCE",axis=1)
heatMapCSV=heatMapCSV.drop("Report Number",axis=1)
heatMapCSV.reset_index()
heatMapCSV
Out[36]:
In [37]:
# Function that looks at the month pulled from crime start date and assigns the numeric month value to its string value
heatMapCSV['Month of OCCURENCE'] = None
for idx in heatMapCSV.index:
curr_month = heatMapCSV.at[idx, "Offense Start DateTime"].month
if curr_month == 12:
heatMapCSV.at[idx,'Month of OCCURENCE'] = "December"
elif curr_month == 11:
heatMapCSV.at[idx,'Month of OCCURENCE'] = "November"
elif curr_month == 10:
heatMapCSV.at[idx,'Month of OCCURENCE'] = "October"
elif curr_month == 9:
heatMapCSV.at[idx,'Month of OCCURENCE'] = "September"
elif curr_month == 8:
heatMapCSV.at[idx,'Month of OCCURENCE'] = "August"
elif curr_month == 7:
heatMapCSV.at[idx,'Month of OCCURENCE'] = "July"
elif curr_month == 6:
heatMapCSV.at[idx,'Month of OCCURENCE'] = "June"
elif curr_month == 5:
heatMapCSV.at[idx,'Month of OCCURENCE'] = "May"
elif curr_month == 4:
heatMapCSV.at[idx,'Month of OCCURENCE'] = "April"
elif curr_month == 3:
heatMapCSV.at[idx,'Month of OCCURENCE'] = "March"
elif curr_month == 2:
heatMapCSV.at[idx,'Month of OCCURENCE'] = "February"
elif curr_month == 1:
heatMapCSV.at[idx,'Month of OCCURENCE'] = "January"
heatMapCSV.head()
Out[37]:
In [38]:
heatPlot=heatMapCSV
heatPlot
Out[38]:
Now to intialize the folium maps. Seattle, WA coordinates on the map are 47.6062° N, 122.3321° W. This translates to [47.6062, -122.331] for folium viewing.¶
In [39]:
# Initial display of Seattle
map_model = folium.Map(location=[47.6062, -122.331], zoom_start=11)
map_model
Out[39]:
In [40]:
# Map to display Offense Parent Group crime locations
map_parentGroup=folium.Map(location=[47.6062, -122.331], zoom_start=11)
map_parentGroup
Out[40]:
In [41]:
# intialize map to display seasons of crime occurances
map_month = folium.Map(location=[47.6062, -122.331], zoom_start=11)
map_month
Out[41]:
In [42]:
# initialize map to display crime against locations
map_crimeAgainst=folium.Map(location=[47.6062, -122.331], zoom_start=11)
map_crimeAgainst
Out[42]:
In [43]:
# In order to plot the Latitude Longitude values of the crime data set, the 2 values must be made into a list
locationsCrime = heatMapCSV[['Latitude', 'Longitude']]
locationsCrime
Out[43]:
In [44]:
# turns LAT,LONG to list
locationsCrimeList=locationsCrime.values.tolist()
locationsCrimeList
Out[44]:
In [45]:
(heatMapCSV['Crime Against Category']).unique()
Out[45]:
In [46]:
(heatMapCSV['Month of OCCURENCE']).unique()
Out[46]:
In [47]:
(heatMapCSV['Offense']).unique()
Out[47]:
In [48]:
# functions checks to see if offense (crime) committed is one of the top 5 most common of all years
def isPopularOffense(offense):
if offense == 'LARCENY-THEFT':
return 'white'
elif offense == 'BURGLARY/BREAKING & ENTERING':
return 'red'
elif offense == 'MOTOR VEHICLE THEFT':
return 'blue'
elif offense == 'DESTRUCTION/DAMAGE/VANDALISM OF PROPERTY':
return 'yellow'
elif offense == 'FRAUD OFFENSES':
return 'green'
In [49]:
# Grabs only the 2020 year values to plot so that crimes in 1 year can be seen
heatPlotIT=heatMapCSV[0:3000]
heatPlotIT
Out[49]:
In [50]:
# function that iterates through 2020 crime data set and plots where any LARCENY-THEFT, BURGLARY/BREAKING & ENTERING,
# MOTOR VEHICLE THEFT, DESTRUCTION/DAMAGE/VANDALISM OF PROPERTY, and FRAUD OFFENSES occured
count_LARCENY_THEFT=0
count_BURGLARY_BREAKING_and_ENTERING=0
count_MOTOR_VEHICLE_THEFT=0
count_DESTRUCTION_DAMAGE_VANDALISM_OF_PROPERTY=0
count_FRAUD_OFFENSES=0
for point in heatPlotIT.index:
if heatPlotIT.loc[point]['Offense Parent Group']=='LARCENY-THEFT':
count_LARCENY_THEFT+=1
elif heatPlotIT.loc[point]['Offense Parent Group']=='BURGLARY/BREAKING&ENTERING':
count_BURGLARY_BREAKING_and_ENTERING+=1
elif heatPlotIT.loc[point]['Offense Parent Group']=='MOTOR VEHICLE THEFT':
count_MOTOR_VEHICLE_THEFT+=1
elif heatPlotIT.loc[point]['Offense Parent Group']=='DESTRUCTION/DAMAGE/VANDALISM OF PROPERTY':
count_DESTRUCTION_DAMAGE_VANDALISM_OF_PROPERTY+=1
elif heatPlotIT.loc[point]['Offense Parent Group']=='FRAUD OFFENSES':
count_FRAUD_OFFENSES+=1
map_parentGroup.add_child(folium.CircleMarker(location=locationsCrimeList[point], fill='true',radius = 6,popup= 'Hi', fill_color=isPopularOffense(heatPlotIT.loc[point]['Offense Parent Group']),color = 'clear',fill_opacity=1))
map_parentGroup
Out[50]:
From this map's display, it is easy to see that the most crimes fall under the Larceny Theft parent offense group (white) for the year 2020, just like for the overall trend for years 2008-2020.¶
In [51]:
# Function that colors the crime against category for crimes committed in 2020
def isCrimeAgainst(group):
if group == 'PROPERTY':
return 'red'
elif group == 'SOCIETY':
return 'green'
elif group == 'PERSON':
return 'blue'
elif group == 'NOT_A_CRIME':
return 'white'
In [52]:
# function that maps the crime against category for crimes committedd in 2020
count_PROPERTY=0
count_SOCIETY=0
count_PERSON=0
count_NotACrime=0
for point in heatPlotIT.index:
if heatPlotIT.loc[point]['Crime Against Category']=='PROPERTY':
count_PROPERTY+=1
elif heatPlotIT.loc[point]['Crime Against Category']=='SOCIETY':
count_SOCIETY+=1
elif heatPlotIT.loc[point]['Crime Against Category']=='PERSON':
count_PERSON+=1
elif heatPlotIT.loc[point]['Crime Against Category']=='NOT_A_CRIME':
count_NotACrime+=1
map_crimeAgainst.add_child(folium.CircleMarker(location=locationsCrimeList[point], fill='true',radius = 6,popup= 'Hi', fill_color=isCrimeAgainst(heatPlotIT.loc[point]['Crime Against Category']),color = 'clear',fill_opacity=1))
map_crimeAgainst
Out[52]:
From this map, it is clear crimes against property (red) were the most common for 2020. This follows the analysis of the overall Seattle Crime data from 2008-2020 because crimes against property were found to be the most common crime.¶
In [53]:
# function that assigns month to seasons to be plotted for crimes committed in 2020
def whatMonth(the_month):
if the_month == 'January':
return 'green'
elif the_month == 'February':
return 'green'
elif the_month == 'March':
return 'yellow'
elif the_month == 'April':
return 'yellow'
elif the_month == 'May':
return 'yellow'
elif the_month == 'June':
return 'white'
elif the_month == 'July':
return 'white'
elif the_month == 'August':
return 'white'
elif the_month == 'September':
return 'blue'
elif the_month == 'October':
return 'blue'
elif the_month == 'November':
return 'blue'
elif the_month == 'December':
return 'green'
In [54]:
# function that maps the season to location for crimes committed in 2020
count_FALL=0
count_WINTER=0
count_SPRING=0
count_SUMMER=0
for point in heatPlotIT.index:
if heatPlotIT.loc[point]['Month of OCCURENCE']=='January':
count_WINTER+=1
elif heatPlotIT.loc[point]['Month of OCCURENCE']=='February':
count_WINTER+=1
elif heatPlotIT.loc[point]['Month of OCCURENCE']=='March':
count_SPRING+=1
elif heatPlotIT.loc[point]['Month of OCCURENCE']=='April':
count_SPRING+=1
elif heatPlotIT.loc[point]['Month of OCCURENCE']=='May':
count_SPRING+=1
elif heatPlotIT.loc[point]['Month of OCCURENCE']=='June':
count_SUMMER+=1
elif heatPlotIT.loc[point]['Month of OCCURENCE']=='July':
count_SUMMER+=1
elif heatPlotIT.loc[point]['Month of OCCURENCE']=='August':
count_SUMMER+=1
elif heatPlotIT.loc[point]['Month of OCCURENCE']=='September':
count_FALL+=1
elif heatPlotIT.loc[point]['Month of OCCURENCE']=='October':
count_FALL+=1
elif heatPlotIT.loc[point]['Month of OCCURENCE']=='November':
count_FALL+=1
elif heatPlotIT.loc[point]['Month of OCCURENCE']=='December':
count_WINTER+=1
map_month.add_child(folium.CircleMarker(location=locationsCrimeList[point], fill='true',radius = 6,popup= 'Hi', fill_color=whatMonth(heatPlotIT.loc[point]['Month of OCCURENCE']),color = 'clear',fill_opacity=1))
map_month
Out[54]:
From this map of crimes occuring per season, it is clear for the year 2020 it was overwhelmingly winter (green) for when most crimes occured. This finding contradicts the overall data from 2008-2020 analysis because for 2008-2020, most crimes were found to occur in the spring, but for 2020's map, we harddly see any crimes committed that occured in the spring (yellow).¶
Conclusion:¶
Crimes in Seattle are at an all time high in the year 2020. The steep rise of crime can be attributed to the rapidly growing cost to live there and thus criminal activity increasing as natives are forced out of their homes to be replaced by wealthy transplants. It is important to note that the most common occuring crimes are non-violent crimes, indicating that some of the rise in crimes (robberies in particular) are committed only out of desperation to survive in an expensive city. Crimes occured the most frequently during spring on average from 2008-2020 but in the year 2020, crimes occured the most during winter. This can be attributed to the fact that shelters are much more limited in the winter and so many impoverished people are left to commit crimes out of necessity and because they have no other resources. I predict Seattle will continue to see it's crimes peak during the winter because of the COVID-19 economic fallout leaving many Americans in poverty.¶
In [ ]: