SP JAIN INSTITUTE OF

MANAGEMENT & RESEARCH

ASSIGNMENT
Operations Management (OM)

Revenue Forecast of Tata Consumer Products

Group Members: (C-3)

Adyasha Das [ PGPM-23-124]
T Sriram Shraff [ PGPM-23-172]
Devarshi Datta [ PGPM-23-137]
Marrapu Sai Chaitanya [ PGPM-23-145]

Tab - Simpson Paradox:

Let us have an aggregate level categorisation of data in High battery and Less battery categories to analyse the
division in the availability of expandable storage.

As observed, the phones with high battery (>4900 mAh) tend to have no expandable storage.
 Now let us analyse at the level of "thickness"

Tab - Probability:

Number of smartphones with OLED display = 30

Number of smartphones with LED display = 77
Total Number of smartphones =107
Therefore, Probability of smartphone being randomly selected to be with OLED display
=
No. of smartphones with OLED display (30)

Total no. of smartphones (107)

This indicates 28.04% of the phones selected randomly are with the OLED display

The likelihood of randomly selecting a smartphone with display refresh rate of 120Hz has been calculated directly by
counting no. of smartphones with display refresh rate of 120Hz and dividing it by total no. of smartphones
Number of smartphones with display refresh rate of 120Hz = 18
Number of smartphones with display refresh rate of 90Hz = 89
Total no. of smartphones =107
Therefore, Probability of smartphone being randomly selected to be with display refresh rate of 120Hz
1
=
No. of No. of smartphones display refresh rate of 120Hz (18)

Total no. of smartphones (107)

This indicates 16.82% of the phones selected randomly are display refresh rate of 120Hz

Conditional Probability method:

Next, they wanted to know that if the company offers LED screens, what is the probability of them offering 90Hz
display refresh rate? For this we will also use probability distribution frequency method. In the pivot table, first we
filter by screen type and then calculate the % of 90 Hz display. We concluded that there is a 89.6% probability that a
phone with LED screen will have an 90Hz screen refresh rate.

For probability of getting both LED screen and 90Hz display, we create a cross-tabulation of display screen type and
screen refresh rate. From that we can conclude that out of the 107 total data points, in 69 instances a selected phone
has both LED screen and 90Hz refresh rate. Hence the probability of getting both LED screen and 90Hz refresh rate is
64.49%.

Bayes Theorem

For their next question of getting the probability of having high camera performance and OLED screen can be solved
using Bayes Theorem. Using Conditional Probability, we get to know that Probability of having OLED and high Camera
performance is 18.78%, while the total probability for getting High Performance camera is 83.55%. Hence probability
of a phone having an OLED display given that the camera is high performing is 22.48%. Below are the snip shots for
the calculation:

For obtaining the probability of High performance camera and OLED Screen, Bayes Theorem is used.

2
The probability of a Phone having OLED Screen with High Performance Camera is 22%.

Tab - Normal Distribution:

By analysing and interpreting the histogram and the mean, median and mode data, it is concluded that the data
follows a Normal Distribution. As mean is nearly equal to mode equal to median.

The Probability of Phones with thickness more than 8.5 mm is 1-NORM.DIST(8.5,8.1, std dev.,TRUE) = 28.52%.

Similarly, The Probability of Phones with thickness less than 8 mm is 45.34%.

The Probability of Phones with thickness in between 7.5mm and 8.5mm is 50.32%.

NORM.INV function in excel is used to calculate the thickness from the given %age.

The thickness of phones with probability of 12% is 7.23 mm.

The thickness of phones with probability of 20% is 7.47 mm.

Tab - Binomial Distribution:

For obtaining the probability of exactly 5 out of 10 batteries to be of above 5000mAh capacity, Binomial Distribution
concept is used.

3
BINOM.DIST function of excel is used.

The Probability of exactly 5 out of 10 batteries to be of above 5000 mAh capacity is 1.35%.

Tab - Hypothesis Testing:

To check the deviation of the average fabrication data from 9mm, Hypothesis Testing is done.

Null and Alternate Hypothesis are formulated.

Null Hypothesis, Ho: Mean is not different from 9

Alternate Hypothesis, Ha: Mean is significantly different from 9

Test statistic of 1.84 is obtained. As p-value is greater than alpha, Null Hypothesis is not rejected.

Hence, it is concluded that the average fabrication data does not significantly deviate from 9mm.

Tab - Confidence Interval:

For obtaining the confidence interval, sigma unknown case is used.

The confidence Interval is ( 8.794, 10.234 ).

4
Tab - 2 Population Hypothesis Testing:

For comparing the prices of Snapdragon and Mediatek Phones, 2 population hypothesis testing with sigma unknown
is used.

Null and Alternate Hypothesis are formulated.

Null Hypothesis, Ho: Mean Prices of Snapdragon and Mediatek phones are equal

Alternate Hypothesis, Ha: Mean Prices of Snapdragon and Mediatek Phones are not equal.

P value is found to be less than alpha value. Hence, the Null Hypothesis of Mean Prices of Snapdragon and Mediatek
phones to be equal is rejected.

5
Test of Independence:

To determine whether the price of phones are influenced by features like Rear / Side Fingerprint Recognition and
Sim-Slot (Dedicated, dual, etc), Test of Independence is used.

Null and Alternate Hypothesis are formulated.

Null Hypothesis, Ho: Mean Prices of Sim slots of dedicated/ dual hybrid are equal

Alternate Hypothesis, Ha: Mean Prices of Sim slots of dedicated/ dual hybrid Phones are not equal.

P value is found to be greater than alpha.

Hence, Null hypothesis is rejected. There exists mean price difference of phones with dedicated/ dual/ hybrid sim
slots.

Again,

Null and Alternate Hypothesis are formulated.

Null Hypothesis, Ho: Mean Prices of Fingerprint Recognition Phones are equal

Alternate Hypothesis, Ha: Mean Prices of Fingerprint Recognition Phones are not equal.

6
P value is found to be greater than alpha.
Hence, Null hypothesis is rejected. There exists mean price difference of phones with Fingerprint Recognition.

Tab - Simple Linear Regression:

To predict the price of phones using weights and fabrication data, Simple Linear Regression Tool is used.

The price is dependent variable and weights and fabrication data are independent variable.

Regression function is used in Excel. The corelation between price and weight is very weak,
corelation between proce and fabrication is also very weak. So,
it is not advisable to use this corelation to predict pricing.

As per the summary output , R square is very low 0.350 and

hence this ascertains our above observation that corelation is
very weak to be used in predicting price.

The corelation between price-weight and price-fabrication is weak. Hence not adviced to use for price prediction.

The value of R-square is also very low. This ascertains our above observation of weak corelation and should not be
used for price prediction.

Tab- Test of Independence:

Case I: Relationship Test between Product Company and SoC (Snapdragon and Mediatek)

Null and Alternate Hypothesis are formulated.

Null Hypothesis, Ho: Product Company and SoC components are independent of each other.

7
Alternate Hypothesis, Ha: Product Company and SoC components are not independent of each other.

Chi- square Test of Independence is carried out.

As p value is less than alpha value, Null Hypothesis is rejected.

So, it is concluded that Product Company and SoC Components are dependent of each other.

Case II: Relationship Test between Product Company and Screen Type

Null and Alternate Hypothesis are formulated.

Null Hypothesis, Ho: Product Company and Screen Type are independent of each other.

Alternate Hypothesis, Ha: Product Company and Screen Type are not independent of each other.

Expected Frequency values are calculated and Chi- square Test of Independence is carried out.

As p value is less than alpha value, Null Hypothesis is rejected.

So, it is concluded that Product Company and Screen Type are dependent of each other.

Case III: Relationship Test between SoC and Fingerprint

Null and Alternate Hypothesis are formulated.

Null Hypothesis, Ho: Fingerprint Sensor position and SoC Components are independent of each other.

Alternate Hypothesis, Ha: Fingerprint Sensor position and SoC Components are not independent of each other.

8
Expected Frequency values are calculated and Chi- square Test of Independence is carried out.

As p value is more than alpha value, Null Hypothesis is not rejected.

So, it is concluded that SoC Type and Fingerprint Sensor Position are independent of each other.