Internet Usage Analysis¶

Answer for the following questions usign Data Analysis.

What is the most frequent internet activity time of the day ?

How often the ip changes ?

How often the device changed.

What is the average usage per hour , per day and per month ?

In [59]:
import numpy as np
import pandas as pd
import os
import warnings
warnings.filterwarnings('ignore')
import matplotlib.pyplot as plt
%matplotlib inline
import seaborn as sns
from sklearn import preprocessing
from sklearn.impute import SimpleImputer
In [60]:
path = 'C:\\Users\\mahit\\OneDrive\\Desktop\\DSPP\\ML & DL\\Assignments\\JNTUH ML DL Assignment 1\\'
In [61]:
df = pd.read_csv(path+'internet_session.csv')
df
Out[61]:
name start_time usage_time IP MAC upload download total_transfer seession_break_reason
0 user1 2022-05-10 02:59:32 00:00:36:28 10.55.14.222 48:E7:DA:58:22:E9 15861.76 333168.64 349030.40 Idle-Timeout
1 user1 2022-05-10 18:53:27 00:01:49:56 10.55.2.253 48:E7:DA:58:22:E9 16957.44 212152.32 229109.76 Idle-Timeout
2 user1 2022-05-10 21:20:44 00:01:35:00 10.55.2.253 48:E7:DA:58:22:E9 14080.0 195153.92 209233.92 Idle-Timeout
3 user1 2022-05-11 00:37:42 00:00:26:00 10.55.2.253 48:E7:DA:58:22:E9 5242.88 40806.4 46049.28 Idle-Timeout
4 user1 2022-05-11 02:59:38 00:00:11:52 10.55.2.253 48:E7:DA:58:22:E9 22067.2 10772.48 32839.68 Idle-Timeout
... ... ... ... ... ... ... ... ... ...
4707 user9 2022-11-04 01:11:34 00:06:54:32 10.55.4.189 DA:2F:97:0E:B7:D0 107960.32 2390753.28 2495610.88 Idle-Timeout
4708 user9 2022-11-04 10:26:09 00:00:23:49 10.55.4.59 DA:2F:97:0E:B7:D0 11407.36 209674.24 221081.60 Idle-Timeout
4709 user9 2022-11-04 20:41:42 00:01:24:13 10.55.15.186 DA:2F:97:0E:B7:D0 18995.2 373657.6 392652.80 Idle-Timeout
4710 user9 2022-11-05 00:21:06 00:08:49:43 10.55.4.159 DA:2F:97:0E:B7:D0 46602.24 593766.4 640368.64 Idle-Timeout
4711 user9 2022-11-05 20:55:37 00:01:06:20 10.55.2.33 DA:2F:97:0E:B7:D0 21237.76 298536.96 319774.72 NaN

4712 rows × 9 columns

In [62]:
df.head()
Out[62]:
name start_time usage_time IP MAC upload download total_transfer seession_break_reason
0 user1 2022-05-10 02:59:32 00:00:36:28 10.55.14.222 48:E7:DA:58:22:E9 15861.76 333168.64 349030.40 Idle-Timeout
1 user1 2022-05-10 18:53:27 00:01:49:56 10.55.2.253 48:E7:DA:58:22:E9 16957.44 212152.32 229109.76 Idle-Timeout
2 user1 2022-05-10 21:20:44 00:01:35:00 10.55.2.253 48:E7:DA:58:22:E9 14080.0 195153.92 209233.92 Idle-Timeout
3 user1 2022-05-11 00:37:42 00:00:26:00 10.55.2.253 48:E7:DA:58:22:E9 5242.88 40806.4 46049.28 Idle-Timeout
4 user1 2022-05-11 02:59:38 00:00:11:52 10.55.2.253 48:E7:DA:58:22:E9 22067.2 10772.48 32839.68 Idle-Timeout
In [63]:
df.isnull().sum()
Out[63]:
name                     0
start_time               0
usage_time               0
IP                       0
MAC                      0
upload                   0
download                 0
total_transfer           0
seession_break_reason    9
dtype: int64
In [64]:
# Data Imputation of seession_break_reason column
imp = SimpleImputer(strategy="most_frequent")
df['seession_break_reason'] = imp.fit_transform(df['seession_break_reason'].values.reshape(-1, 1))
In [65]:
df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 4712 entries, 0 to 4711
Data columns (total 9 columns):
 #   Column                 Non-Null Count  Dtype  
---  ------                 --------------  -----  
 0   name                   4712 non-null   object 
 1   start_time             4712 non-null   object 
 2   usage_time             4712 non-null   object 
 3   IP                     4712 non-null   object 
 4   MAC                    4712 non-null   object 
 5   upload                 4712 non-null   object 
 6   download               4712 non-null   object 
 7   total_transfer         4712 non-null   float64
 8   seession_break_reason  4712 non-null   object 
dtypes: float64(1), object(8)
memory usage: 331.4+ KB
In [66]:
df['name'].value_counts()
Out[66]:
user4    727
user1    674
user6    674
user9    572
user7    528
user3    519
user2    457
user5    336
user8    225
Name: name, dtype: int64
In [67]:
# Data Encoding of MAC using labelEncoder()
label_encoder = preprocessing.LabelEncoder()
df['MAC']= label_encoder.fit_transform(df['MAC'])
df['MAC'].unique()
Out[67]:
array([ 4,  0,  7, 18, 10, 32,  2, 15, 30,  3, 23, 22, 19, 12,  8, 24, 16,
       20, 28,  5, 13, 26, 31, 21,  6,  9, 11, 17, 27, 29,  1, 14, 25])

Exploratory Data Analysis¶

In [68]:
sns.countplot(df['name'])
Out[68]:
<AxesSubplot:xlabel='name', ylabel='count'>
In [69]:
print('The earliest start time is:' )
print(df.start_time.min())
print('The latest start time is:')
print(df.start_time.max())

The earliest start time is:
2022-05-09 22:52:41
The latest start time is:
2022-11-05 21:54:24
In [70]:
print('The minimum usage time is:')
print(df.usage_time.min())
print('The maximum usage time is:')
print(df.usage_time.max())

The minimum usage time is:
00:00:00:01
The maximum usage time is:
01:00:21:07
In [71]:
# Minimum Usage time for each user
minimum_usage_time = df.groupby('name').usage_time.min()
minimum_usage_time
Out[71]:
name
user1    00:00:00:18
user2    00:00:00:08
user3    00:00:00:01
user4    00:00:00:45
user5    00:00:01:07
user6    00:00:00:18
user7    00:00:00:20
user8    00:00:00:20
user9    00:00:00:09
Name: usage_time, dtype: object
In [72]:
# Maximum Usage time for each user
maximum_usage_time = df.groupby('name').usage_time.max()
maximum_usage_time
Out[72]:
name
user1    00:19:35:11
user2    00:20:39:52
user3    00:17:01:28
user4    01:00:21:07
user5    00:06:36:11
user6    00:19:35:11
user7    00:22:00:07
user8    00:17:24:26
user9    00:19:26:09
Name: usage_time, dtype: object

User 7 has the most usage time with 22 hrs 7 sec, User 5 has least usage time with 6 hrs 36 min 11 sec¶

In [73]:
df['IP'].value_counts()
Out[73]:
10.55.0.89      80
10.55.14.148    64
10.55.15.221    55
10.55.1.50      48
10.55.10.46     44
                ..
10.55.14.67      1
10.55.7.44       1
10.55.12.225     1
10.55.12.190     1
10.55.2.33       1
Name: IP, Length: 1302, dtype: int64

Most used IP Address is 10.55.0.89 (80 times)¶

In [74]:
df['MAC'].value_counts()
Out[74]:
4     1236
25     561
12     481
22     465
31     362
24     242
20     237
15     231
1      189
21     147
32     128
0      104
3       73
5       70
19      54
14      36
26      23
17      12
28      12
23      11
2        7
30       7
13       5
10       4
16       3
9        2
6        2
18       2
7        2
11       1
27       1
29       1
8        1
Name: MAC, dtype: int64

Most used device is '4' with 1236 times¶

In [75]:
max_upload = df.groupby('name').upload.min()
max_upload
Out[75]:
name
user1      1006.74
user2     10117.12
user3     10055.68
user4     10147.84
user5     10076.16
user6      1006.74
user7      10035.2
user8     10004.48
user9    100720.64
Name: upload, dtype: object
In [76]:
min_upload = df.groupby('name').upload.max()
min_upload
Out[76]:
name
user1      9932.8
user2     9994.24
user3    99860.48
user4     9963.52
user5     9973.76
user6      9932.8
user7    99860.48
user8     9922.56
user9     9963.52
Name: upload, dtype: object

Highest Max upload is done by User 9 and Lowest Min upload is done by user 8¶

In [77]:
max_download = df.groupby('name').download.max()
max_download
Out[77]:
name
user1    999393.28
user2     99676.16
user3      99635.2
user4      9953.28
user5    990812.16
user6    999393.28
user7    984227.84
user8     97884.16
user9       9984.0
Name: download, dtype: object
In [78]:
min_download = df.groupby('name').download.min()
min_download
Out[78]:
name
user1    100167.68
user2      10035.2
user3     100044.8
user4    100055.04
user5     10076.16
user6    100167.68
user7    100014.08
user8    101672.96
user9    1001216.0
Name: download, dtype: object

Highest Max download is done by User 1 & 6 and Lowest Min download is done by User 2¶

In [79]:
max_total_transfer = df.groupby('name').total_transfer.max()
max_total_transfer
Out[79]:
name
user1    24389877.76
user2    22051553.28
user3     3282042.88
user4     8524922.88
user5     5158993.92
user6    24389877.76
user7    28552724.48
user8     3166699.52
user9     6155141.12
Name: total_transfer, dtype: float64
In [80]:
min_total_transfer = df.groupby('name').total_transfer.min()
min_total_transfer
Out[80]:
name
user1     75.34
user2     73.13
user3      1.12
user4     13.45
user5    924.40
user6     75.34
user7     15.25
user8     18.08
user9    102.64
Name: total_transfer, dtype: float64

Highest Max total transfer is done by User 7 and Lowesrt Min total transfer is done by User 3¶

In [81]:
df['seession_break_reason'].value_counts()
Out[81]:
Idle-Timeout    4359
Lost-Carrier     162
Lost-Service     124
User-Request      65
NAS-Reboot         2
Name: seession_break_reason, dtype: int64
In [82]:
sns.countplot(df['seession_break_reason'])
Out[82]:
<AxesSubplot:xlabel='seession_break_reason', ylabel='count'>
In [83]:
plt.figure(figsize=(18, 9))
ax = sns.countplot(x='name', hue = 'seession_break_reason' , data = df )
ax.bar_label(ax.containers[0])
plt.title("Session break reason per User")
plt.show()
plt.clf()

<Figure size 432x288 with 0 Axes>

The Majority of the session break reasons were from "Idle-Timeout" with 4350 times, while the other reasons have very low occurrences in comparison, with "NAS-Reboot" having the lowest with only 2 occurrences¶

In [ ]: