In [5]:
import matplotlib.pyplot as plt
import matplotlib.style as style
import numpy as np
import pandas as pd
import seaborn as sns
import sklearn.preprocessing as preprocessing
le= preprocessing.LabelEncoder()
import os
#download
import urllib.request as req
import zipfile
import io
In [6]:
from IPython.core.display import HTML as Center
Center(""" <style>
.output_png {
display: table-cell;
text-align: center;
vertical-align: middle;
}
</style> """)
Out[6]:
Downloading¶
In [7]:
#retrieving data
import wget
datasets='/Volumes/Datasets'
project='school_shootings'
url='https://github.com/ecodan/school-shooting-data/archive/refs/heads/master.zip'
zfile=os.path.join(datasets,project,url.split('/')[-1] )
print(zfile)
if not os.path.exists(zfile):
command=('').join(['-P ', os.path.join(datasets, project),' ', url] )
print(command)
!wget $command
else:
print('file exists')
/Volumes/Datasets/school_shootings/master.zip
file exists
In [8]:
print(os.path.join(datasets,project))
/Volumes/Datasets/school_shootings
In [9]:
#unzipping
files_to_extract=['cps_01_formatted.csv', 'pah_wikp_combo.csv']
data_folder=os.path.join(datasets,project,'school-shooting-data-master')
#extract files if they don't exist
for f in files_to_extract:
if not os.path.exists(os.path.join(data_folder,f)):
with zipfile.ZipFile(zfile, 'r') as z:
print(z.namelist())
z.extract(os.path.join('school-shooting-data-master',f), os.path.join(datasets,project))
else:
print(f, 'already exists in folder')
cps_01_formatted.csv already exists in folder
pah_wikp_combo.csv already exists in folder
In [10]:
#loading data into dataframes
def read_csv_into_dataframes(folder):
'''
functions converts .csv files into dataframes and assigns file name as the 'name' attribute for each dataframe that can later be retrieved using df.attrs['name'].
:param folder: a path to .cvs files
:returns: message with names of dataframes loaded
'''
dataframes=[]
print('the following dataframes were read:','\n')
for f in os.listdir(folder):
vars()[f[:-4]]=pd.read_csv(os.path.join(folder,f))
vars()[f[:-4]].attrs['name']=str(f[:-4]) #assigns name of file as name of df to be retrieved later
print(vars()[f[:-4]].attrs['name'])
dataframes.append(vars()[f[:-4]])
return dataframes
In [11]:
dataframes=read_csv_into_dataframes(data_folder)
raw_data_2=dataframes[0]
raw_data_1=dataframes[1]
print('\n\n', raw_data_1.head(3), '\n\n--------------\n\n',raw_data_2.head(3))
the following dataframes were read:
cps_01_formatted
pah_wikp_combo
Date City State AreaType School Fatalities Wounded \
0 3/27/1990 Brooklyn New York NaN C 0 1.0
1 5/20/1990 Centerville Tennessee suburban HS 1 NaN
2 8/26/1990 Las Vegas Nevada urban HS 1 NaN
Dupe Source Desc Unnamed: 10 \
0 NaN Wikp A black youth was taunted with racial slurs by... NaN
1 NaN Pah NaN NaN
2 NaN Pah NaN NaN
Unnamed: 11
0 NaN
1 NaN
2 NaN
--------------
Year Total enrolled N-Total N-Public N-Private K-Total K-Public \
0 2022 72346 3405.0 1888.0 1517.0 3696 3009
1 2021 73456 3801.0 2134.0 1667.0 3765 3145
2 2020 73222 3545.0 2094.0 1451.0 3698 3112
K-Private E-Total E-Public E-Private H-Total H-Public H-Private \
0 687 32005 28300 3705 15957 14747.0 1211.0
1 619 32148 28708 3439 16195 14992.0 1203.0
2 586 32046 28596 3449 16259 14964.0 1296.0
C-Total C-Public C-Private C-Full time
0 17282 13729.0 3553.0 13718.0
1 17547 14084.0 3463.0 13656.0
2 17674 13917.0 3757.0 13774.0
Exploring¶
In [12]:
raw_data_1.info()
<class 'pandas.core.frame.DataFrame'>
RangeIndex: 809 entries, 0 to 808
Data columns (total 12 columns):
# Column Non-Null Count Dtype
--- ------ -------------- -----
0 Date 809 non-null object
1 City 809 non-null object
2 State 809 non-null object
3 AreaType 381 non-null object
4 School 807 non-null object
5 Fatalities 809 non-null int64
6 Wounded 428 non-null float64
7 Dupe 138 non-null object
8 Source 809 non-null object
9 Desc 426 non-null object
10 Unnamed: 10 0 non-null float64
11 Unnamed: 11 0 non-null float64
dtypes: float64(3), int64(1), object(8)
memory usage: 76.0+ KB
Cleaning¶
In [13]:
df=raw_data_1.copy()
In [14]:
#colum_names
def clean_column_names(df):
'''
function to call on after DataFrame is created to format column names in lowercase
without spaces before querying.
:param query: DataFrame
:returns: DataFrame with columns formatted
'''
import pandas as pd
columns=df.columns
new_names=[]
char_dict={' ':'_',':':''}
for c in columns:
new_names.append(c
.lower()
.translate(char_dict)
)
df.columns=new_names
return df
df=clean_column_names(df)
df.info()
<class 'pandas.core.frame.DataFrame'>
RangeIndex: 809 entries, 0 to 808
Data columns (total 12 columns):
# Column Non-Null Count Dtype
--- ------ -------------- -----
0 date 809 non-null object
1 city 809 non-null object
2 state 809 non-null object
3 areatype 381 non-null object
4 school 807 non-null object
5 fatalities 809 non-null int64
6 wounded 428 non-null float64
7 dupe 138 non-null object
8 source 809 non-null object
9 desc 426 non-null object
10 unnamed: 10 0 non-null float64
11 unnamed: 11 0 non-null float64
dtypes: float64(3), int64(1), object(8)
memory usage: 76.0+ KB
In [15]:
#date to timestamp
df.date=pd.to_datetime(df.date)
In [16]:
#school: delete '-" value
df.school=df.school.replace({'-':np.nan})
In [17]:
#state
value_dict={'District Of Columbia':'D.C.', 'District of Columbia':'D.C.', 'IA':'Iowa'}
df.state=df.state.replace(value_dict)
In [18]:
#add calculations
df=(df
.assign(year=df.date.dt.year, month=df.date.dt.month)
.assign(decade=((df.date.dt.year // 10) * 10))
)
df
Out[18]:
| date | city | state | areatype | school | fatalities | wounded | dupe | source | desc | unnamed: 10 | unnamed: 11 | year | month | decade | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | 1990-03-27 | Brooklyn | New York | NaN | C | 0 | 1.0 | NaN | Wikp | A black youth was taunted with racial slurs by... | NaN | NaN | 1990 | 3 | 1990 |
| 1 | 1990-05-20 | Centerville | Tennessee | suburban | HS | 1 | NaN | NaN | Pah | NaN | NaN | NaN | 1990 | 5 | 1990 |
| 2 | 1990-08-26 | Las Vegas | Nevada | urban | HS | 1 | NaN | NaN | Pah | NaN | NaN | NaN | 1990 | 8 | 1990 |
| 3 | 1990-09-11 | San Antonio | Texas | NaN | HS | 0 | 3.0 | NaN | Wikp | 3 students were wounded when gunfire broke out... | NaN | NaN | 1990 | 9 | 1990 |
| 4 | 1991-01-08 | Richardson | Texas | urban | HS | 0 | NaN | NaN | Pah | NaN | NaN | NaN | 1991 | 1 | 1990 |
| ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... |
| 804 | 2023-08-25 | Choctaw | Oklahoma | NaN | HS | 1 | 4.0 | NaN | Wikp | A 16-year-old Midwest City High School student... | NaN | NaN | 2023 | 8 | 2020 |
| 805 | 2023-08-25 | Greensboro | North Carolina | NaN | C | 0 | 1.0 | NaN | Wikp | A non-student was shot on the campus of North ... | NaN | NaN | 2023 | 8 | 2020 |
| 806 | 2023-08-28 | Chapel Hill | North Carolina | NaN | C | 1 | 0.0 | NaN | Wikp | Shooting of Zijie Yan: A graduate student shot... | NaN | NaN | 2023 | 8 | 2020 |
| 807 | 2023-09-12 | Greensburg | Louisiana | NaN | C | 1 | 2.0 | NaN | Wikp | Three people were shot at the St. Helena Colle... | NaN | NaN | 2023 | 9 | 2020 |
| 808 | 2023-09-15 | Country Club Hills | Illinois | NaN | HS | 1 | 0.0 | NaN | Wikp | A 14-year-old boy was fatally shot following a... | NaN | NaN | 2023 | 9 | 2020 |
809 rows × 15 columns
Styling¶
In [19]:
# Color blind friendly colors
# Colorblind-friendly colors
colors = [[0,0,0], [230/255,159/255,0], [86/255,180/255,233/255], [0,158/255,115/255],
[213/255,94/255,0], [0,114/255,178/255]]
style.use('fivethirtyeight')
# sns.axes_style({
# 'axes.facecolor': 'white',
# 'axes.edgecolor': 'black',
# 'axes.grid': False,
# 'figure.facecolor': 'white',
# 'grid.color': 'white',
# 'grid.linestyle': '-',
# 'font.sans-serif': 'Arial',
# 'grid.color': '#ffffff'
# })
footer_left="AAAS24 GIthub - School Schootings"
footer_right="Source: Pah/Amaral/Hagan Research & Wikipedia"
Homework¶
Incidents per Year¶
In [20]:
x_var='rol_decade'
bins=list(range(df.year.max()-30,df.year.max()+1,10))
print(bins)
data=(
df.copy()
.assign(rol_decade=pd.cut(x=df.year,bins=bins, labels=[ '1992-2001', '2002-2011','2012-2022']))
.groupby([x_var])
.agg({'date': "count", 'fatalities':"sum", "wounded":"sum" } )
.assign(total_victims= lambda _df: _df.wounded + _df.fatalities)
.rename(columns={"date":"count_shootings"})
.drop(['fatalities', 'wounded'], axis=1)
.reset_index()
)
data
[1993, 2003, 2013, 2023]
Out[20]:
| rol_decade | count_shootings | total_victims | |
|---|---|---|---|
| 0 | 1992-2001 | 199 | 397.0 |
| 1 | 2002-2011 | 283 | 525.0 |
| 2 | 2012-2022 | 247 | 641.0 |
In [21]:
#incidents per year
title='Shooting Incidents per Decade Over Total Victims (fatalities + wounded)'
subtitle='The past 10 years have seen an increase in victims of gun violence, although a record of incidents remains in the \nprevious decade'
x_var='rol_decade'
num_blank=30 #modify to make footer fit
#data
data=(
df.copy()
.assign(rol_decade=pd.cut(x=df.year,bins=bins, labels=[ '1992-2001', '2002-2011','2012-2022']))
.groupby([x_var])
.agg({'date': "count", 'fatalities':"sum", "wounded":"sum" } )
.assign(total_victims= lambda _df: _df.wounded + _df.fatalities)
.rename(columns={"date":"count_shootings"})
.drop(['fatalities', 'wounded'], axis=1)
.reset_index()
)
data
#
ax1 = sns.set_style(style=None, rc=None )
fig, ax1 = plt.subplots(figsize=(8,16))
##1st chart
y_var='count_shootings'
sns.lineplot(data = data.count_shootings, marker='o', sort = False, ax=ax1, label=y_var)
plt.ylabel(y_var.replace("_"," ").capitalize())
ax2 = ax1.twinx()
##2nd chart
y_var='total_victims'
sns.barplot(data = data, x=x_var, y=y_var, alpha=0.5, ax=ax2, color='#a9a9a9', label=y_var)
plt.ylabel(y_var.replace("_"," ").capitalize())
##labels
plt.axhline(y=0, alpha=0.7, linewidth=1.5, color='black')
##leyend
plt.legend(labelspacing= 2)
sns.set(rc={"figure.figsize":(8 ,16)})
##styling
style.use('fivethirtyeight')
#hiding x label
a=plt.gca()
xax = a.axes.get_xaxis()
xax = xax.set_visible(False)
###Title & Subtitle
plt.text(x=-0.4, y=data[y_var].max()+100,
s=title,
fontsize=26, weight='bold', alpha=.85
)
plt.text(x=-0.4, y=data[y_var].max()+55,
s=subtitle,
fontsize=19, alpha=.85
)
###Footnote Box
plt.tick_params(axis='both', which='major', labelsize=18)
num_spaces=" ".join(" "*int(num_blank))
plt.text(x=-0.4, y=-50,
s=footer_left+num_spaces+footer_right,
fontsize=19,
color="#f0f0f0",
bbox=dict(facecolor="black", alpha=0.5)
)
plt.show()
In [22]:
#incidents per year
title='Shooting Incidents per Decade Over Total Victims (fatalities + wounded)'
subtitle='The past 10 years have seen an increase in victims of gun violence, although a record of incidents remains in the \nprevious decade'
x_var='rol_decade'
num_blank=30 #modify to make footer fit
#data
bins=list(range(df.year.max()-30,df.year.max()+1,10))
data=(
df.copy()
.assign(rol_decade=pd.cut(x=df.year,bins=bins, labels=[ '1992-2001', '2002-2011','2012-2022']))
.groupby([x_var])
.agg({'date': "count", 'fatalities':"sum", "wounded":"sum" } )
.assign(total_victims= lambda _df: _df.wounded + _df.fatalities)
.rename(columns={"date":"count_shootings"})
.drop(['fatalities', 'wounded'], axis=1)
.reset_index()
)
data
#
ax1 = sns.set_style(style=None, rc=None )
fig, ax1 = plt.subplots(figsize=(8,16))
##1st chart
y_var='count_shootings'
sns.lineplot(data = data.count_shootings, marker='o', sort = False, ax=ax1, label=y_var)
plt.ylabel(y_var.replace("_"," ").capitalize())
ax2 = ax1.twinx()
##2nd chart
y_var='total_victims'
sns.barplot(data = data, x=x_var, y=y_var, alpha=0.5, ax=ax2, color='#a9a9a9', label=y_var)
plt.ylabel(y_var.replace("_"," ").capitalize())
##labels
plt.axhline(y=0, alpha=0.7, linewidth=1.5, color='black')
##leyend
plt.legend(labelspacing= 2)
sns.set(rc={"figure.figsize":(8 ,16)})
##styling
style.use('fivethirtyeight')
#hiding x label
a=plt.gca()
xax = a.axes.get_xaxis()
xax = xax.set_visible(False)
###Title & Subtitle
plt.text(x=-0.4, y=data[y_var].max()+100,
s=title,
fontsize=26, weight='bold', alpha=.85
)
plt.text(x=-0.4, y=data[y_var].max()+55,
s=subtitle,
fontsize=19, alpha=.85
)
###Footnote Box
plt.tick_params(axis='both', which='major', labelsize=18)
num_spaces=" ".join(" "*int(num_blank))
plt.text(x=-0.4, y=-50,
s=footer_left+num_spaces+footer_right,
fontsize=17,
color="#f0f0f0",
bbox=dict(facecolor="black", alpha=0.5)
)
plt.show()
In [23]:
#incidents per year
title='Shooting Incidents per Decade Over Total Victims (fatalities + wounded)'
subtitle='Although the number of shooting incidents has not decrease across decades, they have become more deadly'
x_var='rol_decade'
text_x=60 #modify to make footer fit
#data
bins=list(range(df.year.max()-30,df.year.max()+1,10))
data=(
df.copy()
.assign(rol_decade=pd.cut(x=df.year,bins=bins, labels=[ '1992-2001', '2002-2011','2012-2022']))
.groupby([x_var])
.agg({'date': "count", 'fatalities':"sum", "wounded":"sum" } )
.assign(total_victims= lambda _df: _df.wounded + _df.fatalities)
.rename(columns={"date":"count_shootings"})
.drop(['fatalities', 'wounded'], axis=1)
.reset_index()
)
data
ax1 = sns.set_style(style=None, rc=None )
fig, ax1 = plt.subplots(figsize=(8,15))
##1st chart
y_var='count_shootings'
sns.barplot(data = data.loc[:,['rol_decade','count_shootings']], x=x_var, y=y_var, alpha=1, ax=ax1, color='#a9a9a9', label=y_var)
plt.ylabel(y_var.replace("_"," ").capitalize())
plt.tick_params(axis='both', which='major', labelsize=18)
ax2 = ax1.twinx()
##2nd chart
y_var='total_victims'
sns.barplot(data = data.loc[:,['rol_decade','total_victims']], x=x_var, y=y_var, alpha=0.5, ax=ax2, color='red', label=y_var)
plt.ylabel(y_var.replace("_"," ").capitalize())
plt.tick_params(axis='both', which='major', labelsize=18)
##labels
plt.axhline(y=0, alpha=0.7, linewidth=1.5, color='black')
#hiding x label
a=plt.gca()
xax = a.axes.get_xaxis()
xax = xax.set_visible(False)
##leyend
plt.legend(labelspacing= 2)
##styling
style.use('fivethirtyeight')
sns.set_style("white")
###Title & Subtitle
plt.text(x=-0.5, y=data[y_var].max()+90,
s=title,
fontsize=26, weight='bold', alpha=.85
)
plt.text(x=-0.5, y=data[y_var].max()+45,
s=subtitle,
fontsize=19, alpha=.85
)
###Footnote Box
width_ticks=(text_x)
num_spaces=" ".join(" "*int(text_x))
plt.text(x=-0.5, y=-70,
s=footer_left+num_spaces+footer_right,
fontsize=17,
color="#f0f0f0",
bbox=dict(facecolor="black", alpha=0.5)
)
plt.show()
In [ ]:
Evaluating seasonality patterns¶
In [24]:
#fatalities per months
title='Fatalities Registered per Month'
subtitle='Highschool incidents are disproporcially more common in January'
x_var='month'
y_var='count_shootings'
text_x=25
#data
data=(
df.copy()
.groupby(['month', 'school'])
.agg({'date': "count"} )
.rename(columns={"date":"count_shootings"})
.reset_index()
)
data
#plot
##styling
sns.set(rc={"figure.figsize":(15 ,10)}) #(width,height)
plt.tick_params(axis='both', which='major', labelsize=18)
plt.axhline(y=0, alpha=0.7, linewidth=1.5, color='black')
style.use('fivethirtyeight')
# sns.set_style("white")
##plot
ax=sns.lineplot(data=data, x=x_var, y=y_var, hue='school')
##labels
plt.ylabel(y_var.replace("_"," ").capitalize())
plt.xticks(np.arange(1, 13, step=1),['Jan', 'Feb', 'Mar', 'Abr', 'May', 'Jun', 'Jul', 'Ago', 'Sep', 'Oct', 'Nov', 'Dec'] )
#hiding x label
ax.set(xlabel=None)
x=0.1
###Title & Subtitle
plt.text(x=-x, y=data[y_var].max()+10,
s=title,
fontsize=26, weight='bold', alpha=.85
)
plt.text(x=-x, y=data[y_var].max()+5,
s=subtitle,
fontsize=19, alpha=.85
)
###Footnote Box
width_ticks=(text_x)
num_spaces=" ".join(" "*int(text_x))
plt.text(x=-x, y=-10,
s=footer_left+num_spaces+footer_right,
fontsize=17,
color="#f0f0f0",
bbox=dict(facecolor="black", alpha=0.5)
)
plt.show()
Incidents per decade per school type¶
Incidents per decades¶
title='Fatalities per Decades' subtitle='In the two years within the 2020s, we have past the 25% of previous decades, projecting this decade \nto follow or surpass previous ones' x_var='decade' y_var='count_shootings' text_x=20
data¶
data=( df.copy() .groupby([x_var]) .agg({'date': "count"} ) .rename(columns={"date":"count_shootings"}) .reset_index() ) data
plot¶
sns.set_style("white") sns.barplot(data=data, x=x_var, y=y_var, color='grey', alpha=0.5) plt.ylabel(yvar.replace(""," ").capitalize())
hiding x label¶
a=plt.gca() xax = a.axes.get_xaxis() xax = xax.set_visible(False)
style¶
sns.set(rc={"figure.figsize":(15 ,10)}) #(width,height) plt.tick_params(axis='both', which='major', labelsize=18) plt.axhline(y=0, alpha=0.7, linewidth=1.5, color='black') style.use('fivethirtyeight')
x=0.5
Title & Subtitle¶
plt.text(x=-x, y=data[y_var].max()+45, s=title, fontsize=26, weight='bold', alpha=.85 )
plt.text(x=-x, y=data[y_var].max()+25, s=subtitle, fontsize=19, alpha=.85 )
Footnote Box¶
width_ticks=(text_x) num_spaces=" ".join(" "*int(text_x)) plt.text(x=-x, y=-30, s=footer_left+num_spaces+footer_right, fontsize=17, color="#f0f0f0", bbox=dict(facecolor="black", alpha=0.5) )
plt.show()
In [25]:
#Idea: Forcast decade
Fatalities per School Type¶
In [26]:
title='Num Incidents per School Type'
subtitle='Most registered shootings happened in High Schools and Colleges'
x_var='school'
y_var='count_shootings'
text_x=20
data=(
df.copy()
.groupby([x_var])
.agg({'date': "count"} )
.rename(columns={"date":"count_shootings"})
.sort_values("count_shootings", ascending=False)
.reset_index()
)
# data
sns.set(rc={"figure.figsize":(8 ,15)}) #(width,height)
sns.barplot(data=data, x=x_var, y=y_var)
#styling
plt.tick_params(axis='both', which='major', labelsize=18)
plt.axhline(y=0, alpha=0.7, linewidth=1.5, color='black')
##title & subtitle
x=-0.4
plt.text(x=x, y=data[y_var].max()+45,
s=title,
fontsize=26, weight='bold', alpha=.85
)
plt.text(x=x, y=data[y_var].max()+30,
s=subtitle,
fontsize=19, alpha=.85
)
##footer
num_spaces=" ".join(" "*int(text_x))
plt.text(x=x, y=-50,
s=footer_left+num_spaces+footer_right,
fontsize=17, color="#f0f0f0",
bbox=dict(facecolor="black", alpha=0.5)
)
plt.show()
Fatalities per Area Type¶
In [27]:
title='Fatalities per area type'
subtitle='Most registered shootings happened in Urban areas'
x_var='areatype'
y_var='fatalities'
text_x=30
data=(
df.copy()
.loc[:,[x_var,y_var]]
.groupby([x_var]).sum()
.sort_values(y_var, ascending=False)
.reset_index()
)
data
ax=sns.barplot(data=data, x=x_var, y=y_var)
sns.set(rc={"figure.figsize":(8 ,15)}) #(width,height)
#styling
plt.tick_params(axis='both', which='major', labelsize=18)
plt.axhline(y=0, alpha=0.7, linewidth=1.5, color='black')
##hiding x label
ax.set(xlabel=None)
###Title & Subtitle
x=-0.5
plt.text(x=x, y=data[y_var].max()+45,
s=title,
fontsize=26, weight='bold', alpha=.85
)
plt.text(x=x, y=data[y_var].max()+30,
s=subtitle,
fontsize=19, alpha=.85
)
##Footer #1
num_spaces=" ".join(" "*int(text_x))
plt.text(x=x, y=-30,
s=footer_left+num_spaces+footer_right,
fontsize=17, color="#f0f0f0",
bbox=dict(facecolor="black", alpha=0.5)
)
plt.show()
Top 10 states with highest sum of fatalities¶
In [28]:
data=(
df.copy()
.loc[:,['state', 'fatalities']]
.groupby(['state']).sum()
.sort_values('fatalities', ascending=False)
[:10]
)
top_states=data.index.values.tolist()
top_states
Out[28]:
['California',
'Virginia',
'Texas',
'Connecticut',
'Florida',
'Colorado',
'Pennsylvania',
'Washington',
'Tennessee',
'Minnesota']
In [29]:
## Incidents per state
title='Incidents per state'
subtitle='California, Texas and Florida are the states with most frequent shootings'
text_x=25
data=(
df.copy()
[df.state.isin(top_states)]
.loc[:,['year','state']]
.groupby(['state', 'year'])
.agg({'year': "count"} )
.rename(columns={"year":"count_shootings"})
.reset_index()
)
data
# grid
num_states=round(len(top_states))
cols=2
rows=int(np.ceil(num_states/cols))
fix, axes = plt.subplots(rows, cols, figsize=(15,8),sharex=True, sharey=True)
axes = axes.reshape(cols * rows)
for i, k in enumerate(range(0, len(top_states)), 0):
ax = axes[i]
state_i=top_states[i]
state_data=data[data.state==state_i].drop('state', axis=1)
sns.lineplot(data=state_data, x='year', y='count_shootings', ax=ax, label=state_i)
###Title & Subtitle
x=+1945
y=(7.5*5)+(2*5)
plt.text(x=x, y=y+15,
s=title,
fontsize=26, weight='bold', alpha=.85
)
plt.text(x=x, y=y+10,
s=subtitle,
fontsize=19, alpha=.85
)
##Footer #1
num_spaces=" ".join(" "*int(text_x))
plt.text(x=x, y=-7,
s=footer_left+num_spaces+footer_right,
fontsize=17, color="#f0f0f0",
bbox=dict(facecolor="black", alpha=0.5)
)
plt.show()
Fatalities per state (Top 10 states with highest)¶
In [30]:
title='Fatalities per state'
subtitle='California, Texas and Florida incidents seem less deadly that those of other states with less \nincidents, like Virginia & Connecticut'
text_x=25
data=(
df.copy()
.loc[:,['year','state', 'fatalities']]
[df.state.isin(top_states)]
.groupby(['state','year']).sum()
.reset_index()
)
# grid
num_states=round(len(top_states))
cols=2
rows=int(np.ceil(num_states/cols))
fix, axes = plt.subplots(rows, cols, figsize=(15,8),sharex=True, sharey=True)
axes = axes.reshape(cols * rows)
for i, k in enumerate(range(0, len(top_states)), 0):
ax = axes[i]
state_i=top_states[i]
state_data=data[data.state==state_i].drop('state', axis=1)
sns.lineplot(data=state_data, x='year', y='fatalities', ax=ax, label=state_i)
plt.xlabel=(None)
###Title & Subtitle
x=+1945
y=(52*5)+(2*5)
plt.text(x=x, y=y+200,
s=title,
fontsize=26, weight='bold', alpha=.85
)
plt.text(x=x, y=y+150,
s=subtitle,
fontsize=19, alpha=.85
)
##Footer #1
num_spaces=" ".join(" "*int(text_x))
plt.text(x=x, y=-60,
s=footer_left+num_spaces+footer_right,
fontsize=17, color="#f0f0f0",
bbox=dict(facecolor="black", alpha=0.5)
)
plt.show()
ML: Word processing¶
In [31]:
## NPL to Desc: Wordcloud of all data
# libraries for NLP
import nltk
nltk.download('omw-1.4')
from nltk import FreqDist
nltk.download('stopwords')
nltk.download('averaged_perceptron_tagger')
nltk.download('wordnet')
import string
from nltk import pos_tag
from nltk.corpus import stopwords
from nltk.tokenize import WhitespaceTokenizer
from nltk.stem import WordNetLemmatizer
# return the wordnet object value corresponding to the POS tag
from nltk.corpus import wordnet
def get_wordnet_pos(pos_tag):
if pos_tag.startswith('J'):
return wordnet.ADJ
elif pos_tag.startswith('V'):
return wordnet.VERB
elif pos_tag.startswith('N'):
return wordnet.NOUN
elif pos_tag.startswith('R'):
return wordnet.ADV
else:
return wordnet.NOUN
[nltk_data] Downloading package omw-1.4 to
[nltk_data] /Users/alialvarez/nltk_data...
[nltk_data] Package omw-1.4 is already up-to-date!
[nltk_data] Downloading package stopwords to
[nltk_data] /Users/alialvarez/nltk_data...
[nltk_data] Package stopwords is already up-to-date!
[nltk_data] Downloading package averaged_perceptron_tagger to
[nltk_data] /Users/alialvarez/nltk_data...
[nltk_data] Package averaged_perceptron_tagger is already up-to-
[nltk_data] date!
[nltk_data] Downloading package wordnet to
[nltk_data] /Users/alialvarez/nltk_data...
[nltk_data] Package wordnet is already up-to-date!
In [32]:
#define original data
dirty = (
df.copy()
['desc']
.reset_index()
.dropna()
)
dirty
Out[32]:
| index | desc | |
|---|---|---|
| 0 | 0 | A black youth was taunted with racial slurs by... |
| 3 | 3 | 3 students were wounded when gunfire broke out... |
| 5 | 5 | 15-year-old Raul Calvo shot himself dead while... |
| 7 | 7 | A teenager aimed and fired a handgun at a secu... |
| 9 | 9 | 17-year-old Arthur Jermel Jack, was killed by ... |
| ... | ... | ... |
| 804 | 804 | A 16-year-old Midwest City High School student... |
| 805 | 805 | A non-student was shot on the campus of North ... |
| 806 | 806 | Shooting of Zijie Yan: A graduate student shot... |
| 807 | 807 | Three people were shot at the St. Helena Colle... |
| 808 | 808 | A 14-year-old boy was fatally shot following a... |
426 rows × 2 columns
In [33]:
# clean text data
def clean_text(text):
# lower text
text = text.lower()
# tokenize text and remove puncutation
text = [word.strip(string.punctuation) for word in text.split(" ")]
# remove words that contain numbers
text = [word for word in text if not any(c.isdigit() for c in word)]
# remove stop words
stop = stopwords.words('english')
text = [x for x in text if x not in stop]
# remove empty tokens
text = [t for t in text if len(t) > 0]
# pos tag text
pos_tags = pos_tag(text)
# lemmatize text
text = [WordNetLemmatizer().lemmatize(t[0], get_wordnet_pos(t[1])) for t in pos_tags]
# remove words with only one letter
text = [t for t in text if len(t) > 1]
# join all
text = " ".join(text)
return(text)
#define original data
dirty = (
df.copy()
['desc']
.reset_index()
.dropna()
)
##clean data
content_clean=dirty.desc.apply(lambda x: clean_text(x))
##graphic wordcloud
from wordcloud import WordCloud
import matplotlib.pyplot as plt
def show_wordcloud(data, title = None):
wordcloud = WordCloud(
background_color = 'white',
max_words = 200,
max_font_size = 40,
scale = 3,
random_state = 42
).generate(str(data))
fig = plt.figure(1, figsize = (20, 20))
plt.axis('off')
if title:
fig.suptitle(title, fontsize = 20)
fig.subplots_adjust(top = 2.3)
plt.imshow(wordcloud)
plt.show()
show_wordcloud(content_clean)
In [ ]:
In [ ]:
In [ ]: