DATA201_projects/python_project_2.ipynb

In [1]:

import pandas as pd
pd.set_option('display.max_columns', None)

In [2]:

df = pd.read_csv('~/Downloads/NYPD_Complaint_Data_Historic.csv')

/tmp/ipykernel_18634/2268714868.py:1: DtypeWarning: Columns (18,20) have mixed types. Specify dtype option on import or set low_memory=False.
  df = pd.read_csv('~/Downloads/NYPD_Complaint_Data_Historic.csv')

In [3]:

df.columns
# df.dtypes
# df.shape

Out [3]:

Index(['CMPLNT_NUM', 'CMPLNT_FR_DT', 'CMPLNT_FR_TM', 'CMPLNT_TO_DT',
       'CMPLNT_TO_TM', 'ADDR_PCT_CD', 'RPT_DT', 'KY_CD', 'OFNS_DESC', 'PD_CD',
       'PD_DESC', 'CRM_ATPT_CPTD_CD', 'LAW_CAT_CD', 'BORO_NM',
       'LOC_OF_OCCUR_DESC', 'PREM_TYP_DESC', 'JURIS_DESC', 'JURISDICTION_CODE',
       'PARKS_NM', 'HADEVELOPT', 'HOUSING_PSA', 'X_COORD_CD', 'Y_COORD_CD',
       'SUSP_AGE_GROUP', 'SUSP_RACE', 'SUSP_SEX', 'TRANSIT_DISTRICT',
       'Latitude', 'Longitude', 'Lat_Lon', 'PATROL_BORO', 'STATION_NAME',
       'VIC_AGE_GROUP', 'VIC_RACE', 'VIC_SEX'],
      dtype='object')

In [4]:

# According to the data dictionary, CMPLNT_NUM (Complaint Number) is randomly generated and persistent.
# Is it unique?

df['CMPLNT_NUM'].is_unique

Out [4]:

False

In [5]:

# It's not unique. That's unexpected.

df.duplicated(subset = 'CMPLNT_NUM').value_counts()

Out [5]:

False    7821537
True        3962
dtype: int64

In [6]:

# Since CMPLNT_NUM is not unique, we can't use it as an index.
# Let's drop it.

df.drop('CMPLNT_NUM', axis = 1, inplace = True)

In [7]:

df.head(3)
# df.columns
# df.dtypes

Out [7]:

	CMPLNT_FR_DT	CMPLNT_FR_TM	CMPLNT_TO_DT	CMPLNT_TO_TM	ADDR_PCT_CD	RPT_DT	KY_CD	OFNS_DESC	PD_CD	PD_DESC	CRM_ATPT_CPTD_CD	LAW_CAT_CD	BORO_NM	LOC_OF_OCCUR_DESC	PREM_TYP_DESC	JURIS_DESC	JURISDICTION_CODE	PARKS_NM	HADEVELOPT	HOUSING_PSA	X_COORD_CD	Y_COORD_CD	SUSP_AGE_GROUP	SUSP_RACE	SUSP_SEX	TRANSIT_DISTRICT	Latitude	Longitude	Lat_Lon	PATROL_BORO	STATION_NAME	VIC_AGE_GROUP	VIC_RACE	VIC_SEX
0	12/31/2019	17:30:00	NaN	NaN	32.0	12/31/2019	118	DANGEROUS WEAPONS	793.0	WEAPONS POSSESSION 3	COMPLETED	FELONY	MANHATTAN	NaN	STREET	N.Y. POLICE DEPT	0.0	NaN	NaN	NaN	999937.0	238365.0	NaN	NaN	NaN	NaN	40.820927	-73.943324	(40.82092679700002, -73.94332421899996)	PATROL BORO MAN NORTH	NaN	UNKNOWN	UNKNOWN	E
1	12/29/2019	16:31:00	12/29/2019	16:54:00	47.0	12/29/2019	113	FORGERY	729.0	FORGERY,ETC.,UNCLASSIFIED-FELO	COMPLETED	FELONY	BRONX	NaN	STREET	N.Y. POLICE DEPT	0.0	NaN	NaN	NaN	1022508.0	261990.0	NaN	NaN	NaN	NaN	40.885701	-73.861640	(40.885701406000074, -73.86164032499995)	PATROL BORO BRONX	NaN	UNKNOWN	UNKNOWN	E
2	12/15/2019	18:45:00	NaN	NaN	109.0	12/29/2019	578	HARRASSMENT 2	638.0	HARASSMENT,SUBD 3,4,5	COMPLETED	VIOLATION	QUEENS	FRONT OF	STREET	N.Y. POLICE DEPT	0.0	NaN	NaN	NaN	1034178.0	209758.0	25-44	UNKNOWN	M	NaN	40.742281	-73.819824	(40.74228115600005, -73.81982408)	PATROL BORO QUEENS NORTH	NaN	25-44	WHITE HISPANIC	F

In [8]:

# Question 1

# How does each borough compare according to the class of crime committed?
# But first, rename some columns to make the table more readable.

df.rename(columns = {'LAW_CAT_CD': 'CRIME_CLASS', 'BORO_NM': 'BORO'}, inplace = True)
df.groupby(['BORO'])['CRIME_CLASS'].value_counts()

Out [8]:

BORO           CRIME_CLASS
BRONX          MISDEMEANOR    1000078
               FELONY          466248
               VIOLATION       227655
BROOKLYN       MISDEMEANOR    1249836
               FELONY          754414
               VIOLATION       308893
MANHATTAN      MISDEMEANOR    1075687
               FELONY          597184
               VIOLATION       209421
QUEENS         MISDEMEANOR     826883
               FELONY          516528
               VIOLATION       218301
STATEN ISLAND  MISDEMEANOR     210270
               FELONY           81032
               VIOLATION        70589
Name: CRIME_CLASS, dtype: int64

In [9]:

# A quick visualization of the above

df.groupby(['BORO'])['CRIME_CLASS'].value_counts().plot(kind = 'bar')

Out [9]:

<AxesSubplot:xlabel='BORO,CRIME_CLASS'>

In [10]:

# Question 2

# Some incidents occurred in NYC parks, playgrounds or greenspaces.
# What crimes were reported most often and where?

# Again, let's begin by renaming columns.

# df.rename(columns = {'PARKS_NM': 'PUBLIC_SPACE', '', inplace = True}