This is an analysis of some 911 call data from Kaggle that I took as a progress milestone to cover for the first batch of learnings.

Data and Setup

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
%matplotlib inline

Getting started - Data Access

df = pd.read_csv('911.csv')

df.head(3)

df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 99492 entries, 0 to 99491
Data columns (total 9 columns):
lat          99492 non-null float64
lng          99492 non-null float64
desc         99492 non-null object
zip          86637 non-null float64
title        99492 non-null object
timeStamp    99492 non-null object
twp          99449 non-null object
addr         98973 non-null object
e            99492 non-null int64
dtypes: float64(3), int64(1), object(5)
memory usage: 6.8+ MB

Top 5 zipcodes for 911 calls

df['zip'].value_counts().head(5)

19401.0    6979
19464.0    6643
19403.0    4854
19446.0    4748
19406.0    3174
Name: zip, dtype: int64

Top 5 townships (twp) for 911 calls

df['twp'].value_counts().head(5)

LOWER MERION    8443
ABINGTON        5977
NORRISTOWN      5890
UPPER MERION    5227
CHELTENHAM      4575
Name: twp, dtype: int64

How many unique ‘title’ codes are there?

len(df['title'].value_counts())

110

Creating new features

In the titles column there are “Reasons/Departments” specified before the title code.

These are EMS, Fire, and Traffic.

Use .apply() with a custom lambda expression to create a new column called “Reason” that contains this string value.

For example, if the title column value is EMS: BACK PAINS/INJURY , the Reason column value would be EMS.

df['Reason'] = df['title'].apply(lambda title: title.split(':')[0])

df['Reason'].head()

0     EMS
1     EMS
2    Fire
3     EMS
4     EMS
Name: Reason, dtype: object

The most common Reason for a 911 call based off of this new column

df['Reason'].value_counts()

EMS        48877
Traffic    35695
Fire       14920
Name: Reason, dtype: int64

Use Seaborn to Create a countplot of 911 calls by Reason.

sns.countplot(x='Reason', data= df, palette='coolwarm')

Focus on time information.

Data type of the objects in the timeStamp column

type(df['timeStamp'][0])

pandas._libs.tslib.Timestamp

convert the ‘timeStamp’ column from strings to DateTime objects

df['timeStamp'] = pd.to_datetime(df['timeStamp'])

type(df['timeStamp'][0])

pandas._libs.tslib.Timestamp

df['timeStamp'][5]

Timestamp('2015-12-10 17:40:01')

Extract time components eg. Month, hour

print(df['timeStamp'][5].month)
print(df['timeStamp'][5].hour)

12
17

Create 3 new columns called Hour, Month, and Day_of_Week

df['Month'] = df['timeStamp'].apply(lambda timestamp: timestamp.month)

df['Hour'] = df['timeStamp'].apply(lambda timestamp: timestamp.hour)

df['Day_of_week'] = df['timeStamp'].apply(lambda timestamp: timestamp.weekday())

Use the .map() with this dictionary to map the actual string names to the day of the week:

dmap = {0:'Mon',1:'Tue',2:'Wed',3:'Thu',4:'Fri',5:'Sat',6:'Sun'}

df['Day_of_week'] = df['timeStamp'].apply(lambda timestamp: timestamp.weekday()).map(dmap)

With Seaborn create:

Countplot of the Day of Week with the hue based off of the Reason

sns.countplot(data=df, x='Day_of_week',hue='Reason')
plt.legend(loc='center left', bbox_to_anchor=(1.0, 0.5))

Now Countplot for Month:

sns.countplot(data=df, x='Month',hue='Reason')
plt.legend(loc='center left', bbox_to_anchor=(1.0, 0.5))

The dataset is missing some Months,

fill in this information by possibly a simple line plot that fills in the missing months.

Now create a gropuby object called byMonth, where you group the DataFrame by the month column and use the count() method for aggregation.

byMonth = df.groupby('Month').count()

A simple plot of the DataFrame indicating the count of calls per month.

byMonth['twp'].plot()

Create a linear fit on the number of calls per month.

Keep in mind you may need to reset the index to a column.

sns.lmplot(data=byMonth.reset_index(), x='Month',y='twp')

Create a new column called ‘Date’ that contains the date from the timeStamp column.

You’ll need to use apply along with the .date() method.

df['Date'] = df['timeStamp'].apply(lambda timestamp: timestamp.date())

Group by this Date column with the count() aggregate

and create a plot of counts of 911 calls.

df.groupby('Date').count()['twp'].plot()
plt.tight_layout()

Create plots representing a Reason for the 911 call

Plot representing EMS calls

df[df['Reason']=='EMS'].groupby('Date').count()['twp'].plot()
plt.tight_layout()
plt.title('EMS')

Plot representing Fire calls

df[df['Reason']=='Fire'].groupby('Date').count()['twp'].plot()
plt.tight_layout()
plt.title('Fire')

Plot representing Traffic calls

df[df['Reason']=='Traffic'].groupby('Date').count()['twp'].plot()
plt.tight_layout()
plt.title('Traffic')

Creating Heatmaps with Seaborn

Restructure the dataframe so that the columns become the Hours, Index becomes the Day of the Week.

There are lots of ways to do this, try to combine groupby with an unstack method.

dayHour = df.groupby(by=['Day_of_week','Hour']).count()['Reason'].unstack()
dayHour.head()

Create a HeatMap using this new DataFrame

plt.figure(figsize=(12,6))
sns.heatmap(dayHour, cmap='coolwarm')

Create a Clustermap using this DataFrame.

sns.clustermap(dayHour,cmap='viridis')

Repeat these same plots and operations,

for a DataFrame that shows the Month as the column.

dayMonth = df.groupby(by=['Day_of_week','Month']).count()['Reason'].unstack()

Create a HeatMap by Month column

plt.figure(figsize=(12,6))
sns.heatmap(dayMonth, cmap='coolwarm')

Create a Clustermap by Month column

sns.clustermap(dayMonth,cmap='viridis')

So that was the fourth week.. 🔏

https://raw.githubusercontent.com/4bic/4bic.github.io/master/notebooks/911_Project/output_15_1.png