Smartphone Data Analysis¶

Purpose: This notebook analyzes a dataset of smartphones (smartphones_data.csv.csv). The goal is to explore the specifications, features, and pricing trends across different brands. We will perform Exploratory Data Analysis (EDA) to visualize distributions, identify relationships between features, uncover potential insights or anomalies in the data, and summarize the findings. Predictive modeling is not the primary focus but potential challenges for it will be noted.

In [1]:

# Import Libraries
import warnings
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns

# Ignore warnings for cleaner output
warnings.filterwarnings('ignore')

In [2]:

# Load Data
df = pd.read_csv('smartphones_data.csv.csv')

Data Overview and Initial Cleaning¶

In [3]:

# Display basic information
print("Data Info:")
df.info()

# Display first few rows
print("\nData Head:")
df.head()

Data Info:
<class 'pandas.core.frame.DataFrame'>
RangeIndex: 3260 entries, 0 to 3259
Data columns (total 20 columns):
 #   Column                Non-Null Count  Dtype  
---  ------                --------------  -----  
 0   brand_name            3260 non-null   object 
 1   Name                  3260 non-null   object 
 2   Price                 3260 non-null   int64  
 3   RAM                   3260 non-null   float64
 4   OS                    3260 non-null   object 
 5   storage               3260 non-null   float64
 6   Battery_cap           3260 non-null   int64  
 7   has_fast_charging     3260 non-null   object 
 8   has_fingerprints      2534 non-null   object 
 9   has_nfc               2534 non-null   object 
 10  has_5g                2534 non-null   object 
 11  processor_brand       3260 non-null   object 
 12  num_core              3085 non-null   float64
 13  primery_rear_camera   3260 non-null   float64
 14  Num_Rear_Cameras      3260 non-null   int64  
 15  primery_front_camera  3260 non-null   float64
 16  num_front_camera      3260 non-null   int64  
 17  display_size(inch)    3260 non-null   float64
 18  refresh_rate(hz)      1529 non-null   float64
 19  display_types         3260 non-null   object 
dtypes: float64(7), int64(4), object(9)
memory usage: 509.5+ KB

Data Head:

Out[3]:

	brand_name	Name	Price	RAM	OS	storage	Battery_cap	has_fast_charging	has_fingerprints	has_nfc	has_5g	processor_brand	num_core	primery_rear_camera	Num_Rear_Cameras	primery_front_camera	num_front_camera	display_size(inch)	refresh_rate(hz)	display_types
0	vivo	vivo v50	34999	8.0	android	128.0	6000	Yes	Yes	No	Yes	snapdragon	8.0	50.0	2	50.0	1	6.77	120.0	amoled display
1	realme	realme p3 pro	21999	8.0	android	128.0	6000	Yes	Yes	No	Yes	snapdragon	8.0	50.0	2	16.0	1	6.83	120.0	amoled display
2	realme	realme 14 pro plus	27999	8.0	android	128.0	6000	Yes	Yes	No	Yes	snapdragon	8.0	50.0	3	32.0	1	6.83	120.0	oled display
3	samsung	samsung galaxy s25 ultra	129999	12.0	android	256.0	5000	Yes	Yes	Yes	Yes	snapdragon	8.0	200.0	4	12.0	1	6.90	120.0	amoled display
4	vivo	vivo t3 pro	22999	8.0	android	128.0	5500	Yes	Yes	No	Yes	snapdragon	8.0	50.0	2	16.0	1	6.77	120.0	amoled display

In [4]:

# Clean column names (remove special chars, spaces)
df.columns = df.columns.str.replace(r'[()%,]', '', regex=True) # Remove (), %
df.columns = df.columns.str.replace(' ', '_', regex=False) # Replace space with underscore
df.columns = df.columns.str.lower() # Convert to lowercase

# Display cleaned column names
print("Cleaned Column Names:")
print(df.columns)

Cleaned Column Names:
Index(['brand_name', 'name', 'price', 'ram', 'os', 'storage', 'battery_cap',
       'has_fast_charging', 'has_fingerprints', 'has_nfc', 'has_5g',
       'processor_brand', 'num_core', 'primery_rear_camera',
       'num_rear_cameras', 'primery_front_camera', 'num_front_camera',
       'display_sizeinch', 'refresh_ratehz', 'display_types'],
      dtype='object')

In [5]:

# Handle inconsistent brand names
df['brand_name'] = df['brand_name'].replace('moto', 'motorola')

# Convert Yes/No columns to numeric (1/0)
yes_no_cols = ['has_fast_charging', 'has_fingerprints', 'has_nfc', 'has_5g']
for col in yes_no_cols:
    if col in df.columns:
        df[col] = df[col].map({'Yes': 1, 'No': 0})

# Check missing values
print("\nMissing Values Before Handling:")
print(df.isnull().sum())

# Handle missing numerical values with median
num_cols_with_na = ['ram', 'storage', 'num_core', 'primery_rear_camera', 'num_rear_cameras', 
                    'primery_front_camera', 'num_front_camera', 'display_size_inch', 'refresh_rate_hz', 'battery_cap']
for col in num_cols_with_na:
    if col in df.columns:
        # Convert column to numeric, coercing errors
        df[col] = pd.to_numeric(df[col], errors='coerce')
        median_val = df[col].median()
        df[col].fillna(median_val, inplace=True)

# Special handling for Google Pixel processor/cores
df.loc[(df['brand_name'] == 'google') & (df['processor_brand'].isnull()), 'processor_brand'] = 'google'
df.loc[(df['brand_name'] == 'google') & (df['num_core'].isnull()), 'num_core'] = 8 # Assume 8 cores for Tensor

# Handle remaining missing categorical values with 'Unknown'
cat_cols_with_na = ['os', 'processor_brand', 'display_types']
for col in cat_cols_with_na:
     if col in df.columns:
        df[col].fillna('Unknown', inplace=True)

# Fill remaining boolean-like NaNs with 0 (assuming 'No' or feature absent)
for col in yes_no_cols:
    if col in df.columns:
        df[col].fillna(0, inplace=True)

# Convert relevant columns to appropriate types after filling NaNs
for col in num_cols_with_na + yes_no_cols + ['price']:
    if col in df.columns:
         df[col] = pd.to_numeric(df[col], errors='coerce') # Ensure numeric

# Re-check missing values
print("\nMissing Values After Handling:")
print(df.isnull().sum())

# Display cleaned data head
print("\nCleaned Data Head:")
df.head()

Missing Values Before Handling:
brand_name                 0
name                       0
price                      0
ram                        0
os                         0
storage                    0
battery_cap                0
has_fast_charging          0
has_fingerprints         726
has_nfc                  726
has_5g                   726
processor_brand            0
num_core                 175
primery_rear_camera        0
num_rear_cameras           0
primery_front_camera       0
num_front_camera           0
display_sizeinch           0
refresh_ratehz          1731
display_types              0
dtype: int64

Missing Values After Handling:
brand_name                 0
name                       0
price                      0
ram                        0
os                         0
storage                    0
battery_cap                0
has_fast_charging          0
has_fingerprints           0
has_nfc                    0
has_5g                     0
processor_brand            0
num_core                   0
primery_rear_camera        0
num_rear_cameras           0
primery_front_camera       0
num_front_camera           0
display_sizeinch           0
refresh_ratehz          1731
display_types              0
dtype: int64

Cleaned Data Head:

Out[5]:

	brand_name	name	price	ram	os	storage	battery_cap	has_fast_charging	has_fingerprints	has_nfc	has_5g	processor_brand	num_core	primery_rear_camera	num_rear_cameras	primery_front_camera	num_front_camera	display_sizeinch	refresh_ratehz	display_types
0	vivo	vivo v50	34999	8.0	android	128.0	6000	1	1.0	0.0	1.0	snapdragon	8.0	50.0	2	50.0	1	6.77	120.0	amoled display
1	realme	realme p3 pro	21999	8.0	android	128.0	6000	1	1.0	0.0	1.0	snapdragon	8.0	50.0	2	16.0	1	6.83	120.0	amoled display
2	realme	realme 14 pro plus	27999	8.0	android	128.0	6000	1	1.0	0.0	1.0	snapdragon	8.0	50.0	3	32.0	1	6.83	120.0	oled display
3	samsung	samsung galaxy s25 ultra	129999	12.0	android	256.0	5000	1	1.0	1.0	1.0	snapdragon	8.0	200.0	4	12.0	1	6.90	120.0	amoled display
4	vivo	vivo t3 pro	22999	8.0	android	128.0	5500	1	1.0	0.0	1.0	snapdragon	8.0	50.0	2	16.0	1	6.77	120.0	amoled display

Descriptive Statistics¶

In [6]:

# Display descriptive statistics for numerical features
print("Descriptive Statistics (Numerical):")
df.describe()

Descriptive Statistics (Numerical):

Out[6]:

	price	ram	storage	battery_cap	has_fast_charging	has_fingerprints	has_nfc	has_5g	num_core	primery_rear_camera	num_rear_cameras	primery_front_camera	num_front_camera	display_sizeinch	refresh_ratehz
count	3260.000000	3260.000000	3260.000000	3260.000000	3260.000000	3260.000000	3260.000000	3260.000000	3260.000000	3260.000000	3260.000000	3260.000000	3260.000000	3260.000000	1529.000000
mean	20181.384356	5.065874	112.040893	4163.485583	0.473313	0.733436	0.256135	0.307669	7.138037	32.655828	2.076994	12.555767	1.026994	6.097110	100.375409
std	24145.388368	3.256896	126.893532	1312.404904	0.499364	0.442231	0.436564	0.461599	1.649559	29.397695	0.990856	10.564795	0.162090	0.741478	24.920299
min	2500.000000	0.250000	0.310000	1100.000000	0.000000	0.000000	0.000000	0.000000	1.000000	0.300000	1.000000	0.300000	1.000000	2.400000	60.000000
25%	7490.000000	3.000000	32.000000	3007.500000	0.000000	0.000000	0.000000	0.000000	8.000000	12.000000	1.000000	5.000000	1.000000	5.500000	90.000000
50%	11999.000000	4.000000	64.000000	4500.000000	0.000000	1.000000	0.000000	0.000000	8.000000	16.000000	2.000000	8.000000	1.000000	6.455000	120.000000
75%	21999.000000	8.000000	128.000000	5000.000000	1.000000	1.000000	1.000000	1.000000	8.000000	50.000000	3.000000	16.000000	1.000000	6.670000	120.000000
max	200999.000000	24.000000	1024.000000	22000.000000	1.000000	1.000000	1.000000	1.000000	10.000000	200.000000	5.000000	60.000000	2.000000	8.030000	165.000000

In [7]:

# Display descriptive statistics for categorical features
print("\nDescriptive Statistics (Categorical):")
df.describe(include='object')

Descriptive Statistics (Categorical):

Out[7]:

	brand_name	name	os	processor_brand	display_types
count	3260	3260	3260	3260	3260
unique	32	3260	3	15	5
top	samsung	vivo v50	android	mediatek	lcd display
freq	315	1	3130	1384	2010

Exploratory Data Analysis (EDA) - Visualizations¶

Price Distribution¶

In [8]:

plt.figure(figsize=(12, 6))
sns.histplot(df['price'], kde=True, bins=50)
plt.title('Distribution of Smartphone Prices')
plt.xlabel('Price (INR)')
plt.ylabel('Frequency')
plt.show()

plt.figure(figsize=(12, 4))
sns.boxplot(x=df['price'])
plt.title('Box Plot of Smartphone Prices')
plt.xlabel('Price (INR)')
plt.show()

No description has been provided for this image

The price distribution is heavily right-skewed, indicating most phones are in the lower to mid-range, with a few high-priced outliers (likely premium flagships).

Brand Distribution¶

In [9]:

plt.figure(figsize=(14, 8))
sns.countplot(y=df['brand_name'], order = df['brand_name'].value_counts().index)
plt.title('Number of Smartphones per Brand')
plt.xlabel('Count')
plt.ylabel('Brand Name')
plt.tight_layout()
plt.show()

Samsung, Realme, Vivo, Xiaomi, and Oppo are the most represented brands in the dataset.

RAM and Storage Distribution¶

In [10]:

fig, axes = plt.subplots(1, 2, figsize=(15, 5))

sns.countplot(x=df['ram'], ax=axes[0], order = sorted(df['ram'].unique()))
axes[0].set_title('Distribution of RAM (GB)')
axes[0].set_xlabel('RAM (GB)')
axes[0].set_ylabel('Count')

sns.countplot(x=df['storage'], ax=axes[1], order = sorted(df['storage'].unique()))
axes[1].set_title('Distribution of Storage (GB)')
axes[1].set_xlabel('Storage (GB)')
axes[1].set_ylabel('Count')
plt.xticks(rotation=45)

plt.tight_layout()
plt.show()

8GB RAM and 128GB storage are the most common configurations, followed by 6GB/128GB and 12GB/256GB.

Battery Capacity Distribution¶

In [11]:

plt.figure(figsize=(12, 6))
sns.histplot(df['battery_cap'], kde=True, bins=30)
plt.title('Distribution of Battery Capacity (mAh)')
plt.xlabel('Battery Capacity (mAh)')
plt.ylabel('Frequency')
plt.show()

Most phones have batteries around 5000mAh. There are some outliers with very high capacity (> 10000mAh).

Processor Brand Distribution¶

In [12]:

plt.figure(figsize=(10, 6))
sns.countplot(y=df['processor_brand'], order = df['processor_brand'].value_counts().index)
plt.title('Distribution of Processor Brands')
plt.xlabel('Count')
plt.ylabel('Processor Brand')
plt.tight_layout()
plt.show()

Snapdragon and Mediatek dominate the processor market in this dataset.

Feature Presence (5G, NFC, Fast Charging, Fingerprints)¶

In [13]:

fig, axes = plt.subplots(2, 2, figsize=(12, 10))
fig.suptitle('Presence of Key Features (1 = Yes, 0 = No)')

sns.countplot(x=df['has_5g'], ax=axes[0, 0])
axes[0, 0].set_title('5G Support')
axes[0, 0].set_xlabel('Has 5G')

sns.countplot(x=df['has_nfc'], ax=axes[0, 1])
axes[0, 1].set_title('NFC Support')
axes[0, 1].set_xlabel('Has NFC')

sns.countplot(x=df['has_fast_charging'], ax=axes[1, 0])
axes[1, 0].set_title('Fast Charging Support')
axes[1, 0].set_xlabel('Has Fast Charging')

sns.countplot(x=df['has_fingerprints'], ax=axes[1, 1])
axes[1, 1].set_title('Fingerprint Sensor')
axes[1, 1].set_xlabel('Has Fingerprints')

plt.tight_layout(rect=[0, 0.03, 1, 0.95])
plt.show()

Most phones have 5G, Fast Charging, and Fingerprint sensors. NFC is less common, likely absent in budget models.

Price vs. Brand¶

In [14]:

# Filter out potential extreme low price errors for better visualization
df_filtered_price = df[df['price'] > 3000]

plt.figure(figsize=(15, 8))
sns.boxplot(y=df_filtered_price['brand_name'], x=df_filtered_price['price'], order = df_filtered_price.groupby('brand_name')['price'].median().sort_values().index)
plt.title('Price Distribution by Brand (Price > 3000 INR)')
plt.xlabel('Price (INR)')
plt.ylabel('Brand Name')
plt.xscale('log') # Use log scale due to wide price range
plt.tight_layout()
plt.show()

Apple clearly occupies the highest price segment. Brands like Google, Samsung, OnePlus, and Asus cover mid-range to high-end, while Realme, Poco, Xiaomi, Vivo, Oppo, Motorola cover budget to high-end. Brands like Itel, Lava, Tecno are primarily in the budget segment.

Price vs. RAM and Storage¶

In [15]:

fig, axes = plt.subplots(1, 2, figsize=(16, 6))

sns.boxplot(x=df_filtered_price['ram'], y=df_filtered_price['price'], ax=axes[0])
axes[0].set_title('Price vs. RAM (GB)')
axes[0].set_xlabel('RAM (GB)')
axes[0].set_ylabel('Price (INR)')
axes[0].set_yscale('log')

sns.boxplot(x=df_filtered_price['storage'], y=df_filtered_price['price'], ax=axes[1], order=sorted(df_filtered_price['storage'].unique()))
axes[1].set_title('Price vs. Storage (GB)')
axes[1].set_xlabel('Storage (GB)')
axes[1].set_ylabel('Price (INR)')
axes[1].set_yscale('log')
plt.xticks(rotation=45)

plt.tight_layout()
plt.show()

As expected, higher RAM and Storage configurations generally correspond to higher prices.

Correlation Analysis¶

In [16]:

# Select numerical columns for correlation analysis
corr_cols = ['price', 'ram', 'storage', 'battery_cap', 'has_fast_charging', 'has_fingerprints', 
             'has_nfc', 'has_5g', 'num_core', 'primery_rear_camera', 'num_rear_cameras', 
             'primery_front_camera', 'num_front_camera', 'display_size_inch', 'refresh_rate_hz']

# Ensure all selected columns exist and are numeric
valid_corr_cols = [col for col in corr_cols if col in df.columns and pd.api.types.is_numeric_dtype(df[col])]

correlation_matrix = df[valid_corr_cols].corr()

plt.figure(figsize=(14, 10))
sns.heatmap(correlation_matrix, annot=True, cmap='coolwarm', fmt='.2f', linewidths=.5)
plt.title('Correlation Matrix of Numerical Features')
plt.show()

Observations from Correlation Matrix:

price shows positive correlation with ram, storage, has_nfc, has_5g, primery_front_camera, refresh_rate_hz.
ram and storage are positively correlated with each other and with price.
has_5g is positively correlated with price, ram, storage, has_nfc.
refresh_rate_hz is positively correlated with price, ram, storage.
battery_cap doesn't show a strong correlation with price, suggesting high capacity batteries are available across different price points.

Analysis Results Summary¶

Insights¶

The dataset contains a wide range of smartphones from various brands, dominated by Android OS, with iOS (Apple) representing the high-end segment.
Major brands represented include Samsung, Xiaomi, Realme, Vivo, OnePlus, Oppo, Motorola, Apple, Poco, and iQOO.
There is a strong trend towards 8GB RAM and 128GB/256GB storage configurations across mid-range and high-end phones.
5G connectivity is prevalent across most price points, although several budget or older models lack it.
Fast charging ('has_fast_charging' = 1) is almost a standard feature, missing primarily in very low-end or older devices.
AMOLED and OLED display types dominate the mid-range and premium segments, often paired with a 120Hz refresh rate. LCD displays are common in budget phones.
Snapdragon and Mediatek are the most common processor brands, followed by Samsung (Exynos), Apple, and Google (Tensor). Unisoc appears in budget devices.
Most processors are Octa-core (8 cores). Apple uses Hexa-core (6 cores). Some Samsung models list 10 cores.
Primary rear camera resolution varies significantly, with 50MP being very common. High-end models feature 108MP or 200MP sensors.
Battery capacity typically ranges from 4000mAh to 6000mAh, with 5000mAh being a frequent value.
NFC ('has_nfc' = 1) is common in mid-range to high-end phones but often absent in budget models.
Fingerprint sensors ('has_fingerprints' = 1) are standard, except notably on Apple iPhones which use Face ID.

Anomalies¶

Missing values were observed and handled in columns like 'refresh_rate_hz', 'display_types', 'num_core', and 'processor_brand'. Initial handling used median for numerical and 'Unknown' for categorical, with specific logic for Google Pixels.
Inconsistent brand naming ('moto' vs 'motorola') was standardized to 'motorola'.
Presence of very low-spec phones (e.g., RAM < 4GB, Storage < 64GB, no 5G/Fast Charging) like Reliance JioPhone Prima 2 (0.5GB RAM), iKall models, indicating a mix of modern and older/budget devices.
Outlier battery capacities: Oukitel phones listed with >10000mAh (e.g., 11000mAh, 15600mAh, 22000mAh).
Potential data entry errors in price: Google Pixel 6a listed at 4399, Nothing Phone 2a Plus at 2599, Tecno Spark Go 1 at 72908 - these seem incorrect and were filtered for some visualizations.
Missing 'num_core' for Google Pixel phones was handled by assuming 8 cores based on Tensor chip architecture.
Samsung Galaxy S24 FE, S24 Plus 5G, S24 5G listed with 10 cores, which is unusual for mainstream mobile processors.
Apple iPhones consistently listed with 'has_fingerprints' = 0 (No).
Several phones lack 'has_fast_charging', mostly in the budget segment (e.g., Poco C61, iQOO Z9 lite, Itel models, Lava models, some Samsung/Realme budget models).
OnePlus 11R listed with 18GB RAM, potentially a typo or special edition.
Some entries had missing values for both 'refresh_rate_hz' and 'display_types', often older models (e.g., Realme 3 Pro, Samsung Galaxy Note 10 Plus, Xiaomi Redmi Note 9 Pro). These were filled during cleaning.

Predictions¶

Status: Not Performed
Reason: Predictive analysis (e.g., price prediction based on specs) was not performed. The dataset contains inconsistencies (like the 10-core Samsung entries), potential outliers/errors (extreme prices, battery capacities), and requires significant feature engineering (handling numerous brands, processor types, cleaning display types, potentially creating feature interactions) for meaningful and accurate predictions. Such detailed cleaning and modeling are beyond the scope of this automated exploratory analysis. Exploratory Data Analysis provides more direct insights from the current data state.

Overall Summary¶

The smartphone dataset provides a comprehensive overview of the market, showcasing a wide variety of brands, specifications, and price points. Android phones dominate, with major players like Samsung, Xiaomi, Realme, and Vivo offering devices across segments. Key trends include the standardization of 8GB RAM/128GB+ storage, 5G connectivity, fast charging, and high-refresh-rate AMOLED/OLED displays in mid-to-high-end models. Apple maintains its position in the premium segment with iOS. Budget phones often feature LCD screens, lower RAM/storage, lack 5G/NFC, and sometimes fast charging. The data contains anomalies like missing values, potential typos in pricing/specs, and inconsistencies needing further cleaning for robust predictive modeling, but offers valuable insights into current smartphone features and market segmentation.