Telecommunication Industry Project
Introduction
This project focuses on applying descriptive statistics and logical filtering techniques to a
dataset containing specifications and prices of mobile phones. The objective is to analyze
phone features and recommend suitable models based on user preferences using
statistical thresholds such as mean, median, and quartiles.
Dataset Description
The dataset consists of 50 mobile phones and 17 variables describing their technical
specifications and price.
Variables
• PID – Unique phone identifier
• Blue – Bluetooth support (1 = Yes, 0 = No)
• Wi_Fi – WiFi support (1 = Yes, 0 = No)
• Tch_Scr – Touch screen support
• Ext_Mem – External memory support
• Px_h, Px_w – Screen pixels (height and width)
• Scr_h, Scr_w – Screen height and width (cm)
• Int_Mem – Internal memory (MB)
• Bty_Pwr – Battery power (mAh)
• PC – Primary camera (MP)
• FC – Front camera (MP)
• RAM – RAM (GB)
• Depth – Phone thickness (cm)
• Weight – Weight (grams)
• Price – Selling price (₹)
Task 1: Load and Study the Data
Observations
The dataset contains 50 observations (phones).
There are 17 features including PID.
All variables are complete with no missing values.
Features include both categorical (binary) and numerical data.
Task 2: Logical Conditions for Bluetooth, WiFi, Touch Screen & External Memory
Requirement
Children want phones with:
• Bluetooth support
• WiFi support
• Touch screen
• External memory
Logical Condition (con1)
A phone must satisfy all four features simultaneously.
Observation
These features are binary variables.
Only phones satisfying all four conditions are retained for further analysis.
Task 3: Logical Conditions for Screen Resolution (Px_h, Px_w)
Method
A new feature Px = Px_h × Px_w is created to represent total screen resolution.
The median of Px is chosen as the threshold.
Logical Condition (con2)
Phones with Px ≥ median(Px) are selected.
Observation
Using the median ensures selection of phones with above-average screen clarity.
Mean can be used if a less strict condition is required.
Task 4: Logical Conditions for Screen Size (Scr_h, Scr_w)
Method
A new feature Scr_d (Screen Diagonal) is calculated using the formula:
Scr_d = √(Scr_h² + Scr_w²)
The upper quartile (75th percentile) is chosen as the threshold.
Logical Condition (con3)
Phones with Scr_d ≥ upper quartile are selected.
Observation
This ensures only phones with large and premium screen sizes are chosen.
Task 5: Logical Conditions for Camera Quality (PC, FC)
Method
• Mean values of Primary Camera (PC) and Front Camera (FC) are calculated.
Logical Condition (con4)
• Phones with:
o PC ≥ mean(PC)
o FC ≥ mean(FC)
Observation
• This ensures good camera quality for both photography and selfies.
Task 6: Logical Conditions for Internal Memory, Battery & RAM
Features Considered
• Internal Memory (Int_Mem)
• Battery Power (Bty_Pwr)
• RAM
Threshold
• Mean values are used for all three features.
Logical Condition (con5)
• Phones must satisfy:
o Int_Mem ≥ mean(Int_Mem)
o Bty_Pwr ≥ mean(Bty_Pwr)
o RAM ≥ mean(RAM)
Observation
• Ensures smooth performance, sufficient storage, and long battery life.
Task 7: Logical Conditions for Depth and Weight
Requirement
Children prefer slim and lightweight phones.
Threshold
• Median values of Depth and Weight.
Logical Condition (con6)
• Phones with:
o Depth ≤ median(Depth)
o Weight ≤ median(Weight)
Observation
• Median is chosen to avoid extreme values.
Task 8: Final Phone Selection Using All Conditions
Method
• All logical conditions (con1 to con6) are applied simultaneously.
• Resulting dataset is stored as df1.
• Phones are sorted by Price (ascending order).
Results
✓ After applying all filters, 3 phones satisfy all conditions.
✓ Least expensive model: TVF078Y
✓ Most expensive model: TYS938L
Conclusion
These three models can be confidently recommended to the children based on
performance, usability, and value for money.
Task 9: Variability Analysis of Features
Method
• Variability is measured using the ratio:
o Standard Deviation / Mean
• This normalizes variability across features.
Findings
• Most variable feature: Internal Memory
• Least variable feature: Px_w (horizontal pixels)
• Phone prices show high variability, indicating wide market segmentation.
Insight
• Memory capacity varies significantly across models.
• Price differences may be due to branding, camera quality, or premium features.
Final Conclusion
1. Descriptive statistics were effectively used to analyze mobile phone specifications.
2. Logical filtering helped narrow down suitable phone models.
3. Statistical thresholds such as mean, median, and quartiles improved decision accuracy.
4. Three optimal phone models were recommended based on customer preferences