911 Calls Capstone Project¶
- For this capstone project I will be analyzing the 911 call data from [Kaggle].(https://www.kaggle.com/mchirico/montcoalert)¶
In [1]:
from IPython.display import HTML
HTML('''<script>
code_show=true;
function code_toggle() {
if (code_show){
$('div.input').hide();
} else {
$('div.input').show();
}
code_show = !code_show
}
$( document ).ready(code_toggle);
</script>
The raw code for this IPython notebook is by default hidden for easier reading.
To toggle on/off the raw code, click <a href="javascript:code_toggle()">here</a>.''')
Out[1]:
In [1]:
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
%matplotlib inline
sns.set_style('whitegrid')
plt.style.use(u'ggplot')
from mpl_toolkits.basemap import Basemap
In [2]:
from mpl_toolkits.basemap import Basemap
In [3]:
df = pd.read_csv('911.csv')
In [4]:
df.info()
In [5]:
df.head()
Out[5]:
The top 5 zipcodes for 911 calls¶
In [6]:
df['zip'].value_counts().head(5)
Out[6]:
The top 5 townshup for 911 calls¶
In [7]:
df['twp'].value_counts().head(5)
Out[7]:
The most common reason for 911 calls¶
In [8]:
df['Reason'] = df['title'].apply(lambda title: title.split(':')[0])
df['Reason'].value_counts()
Out[8]:
- From the below barchart, we can tell that the 911 calls mostly comes from emergency accidents and fire reason is the least reason for people to call 911. Regarding to traffic accident, it's in the middle between emergency and fire.
In [9]:
plt.figure(figsize=(8,5))
sns.countplot(x='Reason', data=df, palette='viridis')
plt.title('Different Reasons for 911 Calls')
Out[9]:
In [10]:
df['timeStamp']=pd.to_datetime(df['timeStamp'])
df['Hour'] = df['timeStamp'].apply(lambda time: time.hour)
df['Month'] = df['timeStamp'].apply(lambda time: time.month)
df['Day of Week'] = df['timeStamp'].apply(lambda time: time.dayofweek)
dmap = {0:'Mon',1:'Tue',2:'Wed',3:'Thu',4:'Fri',5:'Sat',6:'Sun'}
df['Day of Week'] = df['Day of Week'].map(dmap)
order = ["Sun", "Mon", "Tue", "Wed", "Thu", "Fri", "Sat"]
The number of the 911 calls during a week¶
- From the graph, we can see that during a week, there is the highest 911 calls happening in a week. We can also find that on weekends, the fewer numbers of the 911 calls occured.
In [11]:
plt.figure(figsize=(8,5))
sns.countplot(x='Day of Week',data=df,palette='viridis', order=order)
plt.title('The Number of the 911 Calls in Days of a Week')
plt.tight_layout()
- Dividing the above graph into different reasons and see the difference for 911 calls among various reasons
In [12]:
plt.figure(figsize=(8,5))
sns.countplot(x='Day of Week',data=df,hue='Reason',palette='viridis', order=order)
plt.legend(bbox_to_anchor=(1.05, 1), loc=2, borderaxespad=0.)
plt.title('The Number of the 911 Calls for Different Reasons in a Week')
plt.tight_layout()
- From the below graph, September, October, and November are missing because of the data collection. Due to some missing data, we will study the existing data to see what month will have the highest 911 calls. From the bar chart, we can see that throughout a year, January has the highest 911 calls and December has the lowest ones. It may be bacause holiday is the vacation month, so there are few business operation and few commuters on the road. Most people may stay home and reunite their family members.
In [13]:
plt.figure(figsize=(8,5))
sns.countplot(x='Month',data=df,palette='viridis')
Out[13]:
- Dividing the above graph into different reasons and see the difference for 911 calls among various reasons
In [14]:
plt.figure(figsize=(8,5))
sns.countplot(x='Month',data=df,hue='Reason',palette='viridis')
plt.legend(bbox_to_anchor=(1.05, 1), loc=2, borderaxespad=0.)
Out[14]:
- From the line graph, the roughly decreasing number of 911 calls throughout the year, but there is a peak at July in this declining trend.
In [15]:
byMonth = df.groupby('Month').count()
byMonth['twp'].plot(color='blue')
Out[15]:
- I created the linear regression plot to see its trend and confidence interval.
In [16]:
sns.lmplot(x='Month',y='twp',data=byMonth.reset_index())
Out[16]:
The heatmap of days of a week and hours¶
- There are a heatmap and a clustermap. From these graphs, we can see that the time from 7am to 6pm has the higher 911 calls. The weekdays also have the higher probability of getting 911 calls.
In [17]:
dayHour = df.groupby(by=['Day of Week','Hour']).count()['Reason'].unstack()
dayHour
Out[17]:
In [19]:
plt.figure(figsize=(10,5))
sns.heatmap(dayHour,cmap='Blues')
Out[19]:
In [70]:
plt.figure(figsize=(6,3))
sns.clustermap(dayHour,cmap='Blues')
Out[70]:
In [22]:
dayMonth = df.groupby(by=['Day of Week', 'Month']).count()['Reason'].unstack()
dayMonth
Out[22]:
In [23]:
plt.figure(figsize=(10,5))
sns.heatmap(dayMonth,cmap='Oranges')
Out[23]:
In [24]:
plt.figure(figsize=(8,4))
sns.clustermap(dayMonth, cmap='Oranges')
Out[24]:
In [20]:
lat = df['lat'].values
lon = df['lng'].values
bygeo = df.groupby(by=['lat','lng']).count()['Reason'].values