Sample phone bill analysis

SOMDEEP SEN; Business Analyst: Trimax Analytics

(e) [email protected]; (p): 09748229123

LinkedIn: http://linkd.in/1ifqs3x

• XYZ Consultants have been the task to analyze a sample bill of Vodafone

• The analysis should contain the following details:

Call:

– Total no. of calls made; Called numbers & destinations

– Duration of calls (Total & Avg.); Time of call (Peak, Off-Peak);Cost of calls; Breakup: Day & Date

wise

SMS:

– Total number of SMS sent; Breakup: Day & Date wise

Data Usage:

– Total amount of usage(in KB), Breakup: Day & Date wise

• Determine the type of Usage (Roaming, Local or International Roaming)

• Review the plan to determine overall fitness of the plan for this person

• Recommend an alternative plan based on the comparison with 3 other plans

• Predict certain demographic details(age, occupation) & type of phone being used

Note: One may justify the present plan as the best one too

Type Rate

Local CallMobile(In-Network & Off-Network), Landline: 0.3/Min.

Free Qty: 400 minutes

STDIn-Network & Off-Network Mobile: Rs. 0.5/Min

Local Landline: Rs. 0.5/Min

Roaming CallsIncoming: Rs. 1.0 / minute; Local: Rs. 1.5/ minute

STD: Rs. 0.5/ minute

SMSLocal SMS: Rs. 1.0/ minute; National SMS: Rs. 1.5/ minute

Roaming SMS (National & Local): Rs. 1.5/ minute

Data GPRS Data & GPRS Roaming (0.1 / 10 KB)

Understanding the Data

Data Cleaning

Analysis

Comparison with other plans

Providing recommendation

Prediction

Note

• The information mentioned above is provided for the billing cycle: 10.12.13 to 09.01.14

• The information scattered over multiple excel sheets

Particulars Details

Call

•Date(DD/MM/YY) , Time, Duration (Min : Sec)

•Called Number (Vodafone or Non-Vodafone)

•Type of call (Local or STD) & Charges

SMS

•Date(DD/MM/YY) , Time

•Destination Number (Vodafone or Non-Vodafone)

•Type of SMS (Local or STD)

Data Usage•Date(DD/MM/YY) , Time

•Type of Usage (VF Mob Connect/ VF Live)

Problems Encountered

• Date & Time for the particulars were given together one cell

• Information regarding Conference Call wasn’t specified

• Format of some of the mobile numbers were not in symmetry

• Data usage was scattered in over 10 sheets

Tool Used for cleaning, organizing & analyzing the data: Ms-Excel

• Advanced Filtering

• Pivot Table

• V-Lookup

• This excludes conference call (40 minutes), as the details of con-call was not provided

0

5

10

15

20

25

10

-Dec

11

-Dec

12

-Dec

13

-Dec

14

-Dec

15

-Dec

16

-Dec

17

-Dec

18

-Dec

19

-Dec

20

-Dec

21

-Dec

22

-Dec

23

-Dec

24

-Dec

25

-Dec

26

-Dec

27

-Dec

28

-Dec

29

-Dec

30

-Dec

31

-Dec

1-J

an

2-J

an

3-J

an

4-J

an

5-J

an

6-J

an

7-J

an

8-J

an

6

16

34 3

5

9

1415

14

2 2

25

18

8

24

20

14

5

1

3

1

23

Du

rati

on

(M

ins)

Date

• Total duration: 235 mins; This includes night speak of 11 mins.

• Average duration: 10 mins/day; 23 days have been considered

• No calls were made on 14th Dec, 16th – 19th Dec, 27th – 28th Dec

• Only night calls were made on 25th & 26th Dec

• Drop in average duration weekends suggests that primary purpose of use professional

• Calculated duration(235 minutes) doesn’t match with the billed duration (335 mins)

• Therefore, either some call details might be missing or it might be a case of over billing

• However, both billed & calculated duration is less than 400 mins

• As per the bill 5 minutes STD call was made; but the data didn’t seem to have the details

• All calls made seemed to be Local

• Almost 96% of the calls just lasted <= 3 minutes

• Short duration of calls reemphasizes that the calls might have been professional in nature

0

20

40

60

80

100

120

140

1 minute calls 2 minute calls 3 minute calls 4 minute calls 5 minute calls

123

31

84

2

Nu

mb

er o

f ca

lls

Call Duration

• ‘10:30 AM-7:30 PM’ can be tagged as the peak time as it contributes 180 mins out of 235

• ‘10:30 AM-1:30 PM’ may also be the office time for the person leading to the steep rise

• Interestingly no calls are made between 12:04 PM to 1:04 as it might be the recess time

• Steady drop in call duration suggests that the person might be on the way back home from office

• The before 10:30 AM & after 8 PM can be tagged as off-peak time

17

28

53

99

26

12

0

20

40

60

80

100

120

7:30 AM-10:30 AM 10:30 AM-1:30 PM 1:30 PM-4:30 PM 4:30 PM-7:30 PM 7:30 PM-10:30 PM After 10:30 PM

Du

rati

on

(M

ins)

Time range

• Local Vodafone number clearly occupies the majority of the duration breakup

• The analysis also showed that majority of the numbers were called multiple times in a day

• This suggests that the calls might have been made to the colleagues

• Others included: Local Landline & Toll-free numbers

• Landline numbers also suggested that calls might have been made to different branches of the organization

78

52

105

In Minutes

Non-Voda_Loc_Mob Others Voda_Loc_Mob

This analysis is useful keeping in mind that Vodafone has to pay a termination fee

to end the call made to a Non-Vodafone number

65

43

6

In Minutes

Airtel

Reliance

Docomo

MTS

0

5

10

15

20

25

30

35

40

45

10

-Dec

11

-Dec

12

-Dec

13

-Dec

14

-Dec

15

-Dec

16

-Dec

17

-Dec

18

-Dec

19

-Dec

20

-Dec

21

-Dec

22

-Dec

23

-Dec

24

-Dec

25

-Dec

26

-Dec

27

-Dec

28

-Dec

29

-Dec

30

-Dec

31

-Dec

1-J

an

2-J

an

3-J

an

4-J

an

5-J

an

6-J

an

7-J

an

8-J

an

14

41

44

6

11

33

86

13

810

28

1

20

1 1

Nu

mb

er

Date

• Total no. of SMS sent: 234; Avg. no of SMS sent/day: 14; Days considered: 17

• All the SMS sent was local

• No SMSs were sent on 14th Dec– 19th Dec & 1st – 6th Jan.

• Interestingly no calls were made on 14th Dec, 16th – 19th Dec & only 3 calls were made on the 15th

• This may infer that the person might have been on leave during 14th- 19th December

• However, despite low call duration from 4th to 6th Dec the person seemed active

• Also just like calls Calculated number (236) doesn’t match with the billed one 294)

• The person had an SMS pack of 100 free SMS FOR Rs. 50

13

18

Average Number of SMS

Weekday Weekend

9.5

10.5

Average Duration of Calls

Weekend Weekdays

• Average duration & number is closely matched for calls; whereas for SMS weekend has won the race

• Such stats may infer that due to workload the person may have to work on weekends too

• Assumption: The person uses the phone for professional purpose only

0

100000

200000

300000

400000

500000

600000

700000

800000

900000

• Total usage: 2.08 GB, Avg. data download/day 0.08 GB

• There were no internet usage on 14th Dec, 16th -19th Dec

• Interestingly during these time periods SMS sent & calls made was also very low

• This indicates that the person might have been on a leave

• Maximum usage: 0.79 GB on22ⁿᵈ Dec followed by 27th,23rd 21stand 20th Dec

• There was a modest use of VF Live leading to a negligible consumption of 70 KB

0

5000

10000

15000

20000

25000

30000

35000

40000

12

:00

AM

12

:34

AM

1:0

5 A

M1

:35

AM

2:0

7 A

M2

:42

AM

3:1

6 A

M3

:51

AM

4:2

4 A

M4

:57

AM

5:3

3 A

M6

:06

AM

6:4

0 A

M7

:05

AM

7:3

0 A

M7

:55

AM

8:2

1 A

M8

:53

AM

9:2

8 A

M9

:58

AM

10

:26

AM

10

:59

AM

11

:27

AM

11

:43

AM

12

:02

PM

12

:19

PM

12

:48

PM

1:1

8 P

M1

:46

PM

2:1

3 P

M2

:43

PM

3:0

9 P

M3

:40

PM

4:0

8 P

M4

:39

PM

5:0

4 P

M5

:32

PM

5:5

6 P

M6

:19

PM

6:4

1 P

M7

:07

PM

7:2

8 P

M7

:51

PM

8:2

0 P

M8

:44

PM

9:0

4 P

M9

:22

PM

9:4

1 P

M1

0:0

8 P

M1

0:3

7 P

M1

1:0

6 P

M1

1:3

6 P

M

• Three significant time slots are observed

– 6:30 AM - 8:00 AM, 11:00 AM – 12:30 PM and 8:30 PM – 10:00 PM

These three time slots probably signify:

The starting of the day with checking mails and leads for planning the day ahead

Assigning work and communicating for work purpose

Provide feedback, reviews and instruction for the next day

0

100000

200000

300000

400000

500000

600000

700000

800000

900000

09

-12

-20

13

10

-12

-20

13

11

-12

-20

13

12

-12

-20

13

13

-12

-20

13

15

-12

-20

13

20

-12

-20

13

21

-12

-20

13

22

-12

-20

13

23

-12

-20

13

24

-12

-20

13

25

-12

-20

13

26

-12

-20

13

27

-12

-20

13

28

-12

-20

13

29

-12

-20

13

30

-12

-20

13

31

-12

-20

13

01

-01

-20

14

02

-01

-20

14

03

-01

-20

14

04

-01

-20

14

05

-01

-20

14

06

-01

-20

14

07

-01

-20

14

08

-01

-20

14

09

-01

-20

14

FALSE

TRUE

• Data usage wasn’t very high on weekends except for 21st and 22ⁿᵈ Dec

• It pulled the total usage during weekends higher than the total usage during weekdays

• We can conclude, that this person was working on a certain weekend due to increase in work-load

1. Occupation: High or Mid Level Executive

Justification:

• Relatively short length, but high frequency of calls

• Steep rise in call durations in 10:30 to 7:30

• Spike in 4:30-7:30 indicates that the calls might be used to take reports or feedback at the end of day

• Closely matched stats for avg. no. of calls & SMS on weekdays & weekends suggest high work-Load

• The person may belong to the domain of marketing & sales

2. Age: Greater than Equal to 30

Justification:

• Usually high level executives belong to the age bracket of 30+

3. Type of phone : Blackberry or Smartphone

Justification:

• The customer had access to internet starting from 6:30 AM to 11:30 PM