Data Analytics using Python (AIL0202)
Course: [Link] (CSE – AIML) | I Year II Semester Under COE
Question Bank
Unit 1 – Data Analytics Lifecycle and Python Ecosystem
Question Marks
Define the Data Analytics Lifecycle and explain the significance of each phase. 1
Critically compare descriptive and prescriptive analytics with relevant real-world 3
examples.
Evaluate the role of Data Analytics in advancing AI and Machine Learning 3
capabilities.
Critically examine the Python ecosystem (NumPy, Pandas, Matplotlib, Seaborn) and 5
analyze how they interact within a data analytics workflow.
Describe the different types of analytics: descriptive, diagnostic, predictive, and 5
prescriptive. Justify each with a suitable real-world example.
What distinguishes diagnostic analytics from descriptive analytics? Provide an 1
example from healthcare.
Explain the role of Python in data science. Why is it preferred over R for large-scale 3
analytics projects?
What is the CRISP-DM model? How does it relate to the Data Analytics Lifecycle? 3
Describe three real-world business problems where predictive analytics has been 3
successfully applied.
Compare structured and unstructured data. Provide two examples of each and 1
discuss analytical challenges.
What is the difference between data analytics and data science? Explain with 1
examples.
Evaluate the importance of the 'Problem Discovery' phase in the Data Analytics 3
Lifecycle.
What is big data analytics? Explain the 5 V's (Volume, Velocity, Variety, Veracity, 3
Value) of big data.
Define KPI (Key Performance Indicator) in analytics. Provide three examples from 1
retail analytics.
Critically evaluate how real-time analytics differs from batch analytics in terms of 5
architecture and use cases.
What is a data warehouse? How does it differ from a data lake in architecture and 1
use case?
Analyze how cloud platforms (AWS, GCP, Azure) support large-scale data analytics 3
workflows.
What is exploratory data analysis (EDA)? Why is it a critical first step in any analytics 1
project?
Describe the role of a data analyst versus a data engineer. How do their 3
responsibilities complement each other?
Critically assess how Matplotlib and Seaborn differ in their capabilities for statistical 5
data visualization.
Unit 2 – NumPy Arrays and Pandas DataFrames
Question Marks
Construct matrix A = [[2,4,6],[1,3,5],[7,8,9]] using NumPy syntax and explain the 1
choice of data structure.
Demonstrate how to extract the second row and third column of matrix A using 3
NumPy slicing. Analyze the indexing mechanism.
Using NumPy, compute the row-wise mean of matrix A = [[2,4,6],[1,3,5],[7,8,9]] 3
and interpret the data distribution.
How does transposing matrix A alter its structure? Compute A × Aᵀ and discuss 5
the properties of the resulting matrix.
Construct a 3×3 random matrix using NumPy and analyze the effect of 5
broadcasting when adding a 1D array [1,2,3].
What is broadcasting in NumPy? Give three examples of valid broadcasting 3
operations.
Explain NumPy array indexing: basic, fancy, and boolean indexing with code 3
examples.
What is the difference between [Link]() and view in NumPy? Why is this 1
distinction important?
Write NumPy code to compute the determinant and inverse of a 3×3 matrix. 3
Interpret the result.
Analyze the computational advantages of NumPy vectorization over Python loops 3
for large-scale numerical operations.
Explain the concept of a DataFrame in Pandas and analyze its structural 1
advantages for data manipulation.
How would you load the [Link] dataset into a Pandas DataFrame and inspect 3
the first three rows?
Derive the total marks for each student in the SPD dataset and interpret what this 3
metric reveals about academic performance.
Apply conditional filtering to isolate Section A students in SPD. What patterns 5
emerge from this subset?
Apply a group-by operation on the Section column to compute average marks. 5
Compare insights between the sections.
Describe the difference between loc[] and iloc[] in Pandas. Provide a practical 1
example for each.
What is a Pandas Series? How does it differ from a DataFrame in terms of 1
structure and use?
How would you merge two DataFrames in Pandas? Explain inner, outer, left, and 3
right joins with examples.
Write Pandas code to sort the SPD dataset by Math marks in descending order. 3
What does the ranking reveal?
Critically evaluate the performance differences between Pandas and NumPy for 5
large dataset operations.
Unit 3 – Data Preprocessing
Question Marks
Examine the ESD dataset and identify a data quality issue. Explain how it could 1
impact analytical reliability.
Evaluate suitable methods to handle missing Salary values in the ESD dataset 3
and justify the most reliable approach.
How would you detect duplicate records in the ESD dataset? Discuss the 3
implications of duplicates on analytical outcomes.
Apply the IQR method to detect salary outliers in the ESD dataset. Evaluate their 5
influence on analysis results.
Critically evaluate the step-wise preprocessing workflow for the ESD dataset, 5
justifying the sequence of each step.
Compare mean imputation, median imputation, for handling missing numerical 3
values. When is each preferred?
What is the difference between label encoding and one-hot encoding? Analyze 3
which is more appropriate for the Department column.
Explain min-max normalization and z-score standardization. How does normalizing 3
Salary improve model fairness?
What is feature engineering? Describe three techniques to create new features 3
from the ESD dataset.
What is an outlier? Compare the Z-score method and IQR method for outlier 1
detection. Which is more robust?
What is the difference between data cleaning and data transformation? Provide an 1
example of each.
How would you handle categorical variables with high cardinality (e.g., City with 3
500 unique values)?
Critically evaluate the complete preprocessing pipeline for a mixed dataset 5
containing numerical, categorical, and missing data.
Unit 4 – Data Visualization and Case Studies
Question Marks
Describe how to construct a histogram of Sales using Matplotlib. What patterns or 3
trends can be inferred?
Construct a scatter plot between Advertising_Spend and Sales. What visual 5
patterns are observed?
Critically investigate the Customer Purchase Dataset (CPD). Describe appropriate 5
cleaning and preprocessing techniques.
Analyze the descriptive statistics of CPD. What do the summary measures reveal 5
about customer purchasing behavior?
How would you construct a bar chart and boxplot to communicate distribution and 5
trends in CPD purchase data?
Compare average purchase amounts between Gold and Silver members in CPD. 3
What conclusions can be drawn?
Based on CPD analysis, evaluate key business insights and recommend 3
strategies to enhance customer engagement.
Preprocess the EPD dataset and explain how each preprocessing step ensures 5
data integrity and readiness.
Examine the correlation among variables in EPD. Which relationships are most 3
significant for predicting performance?
Question Marks
Apply a hypothesis test to evaluate the significance of the relationship between 5
Experience and Salary in EPD. Interpret the results.
What are the principles of Gestalt in data visualization? How do they influence 1
chart design choices?
Compare Seaborn's heatmap and pairplot visualizations. When would you use 3
each for multivariate EDA?
What is a boxplot? Describe its components (median, IQR, whiskers, outliers) and 1
explain what it reveals about salary distribution.
Describe three types of charts (line, bar, pie) and discuss which is most 3
appropriate for visualizing monthly sales trends.
What is the role of interactive visualizations (e.g., Plotly, Bokeh) in data analytics 1
dashboards?
How would you use Seaborn's catplot to compare Sales across Regions in the SD 3
dataset? Interpret the result.
Critically evaluate how the storytelling principle applies to data visualization in 3
executive dashboards.
Formulate actionable managerial recommendations based on the EPD analysis to 5
improve employee performance and compensation strategies.
What is a violin plot? How does it extend the information presented by a boxplot? 1
Critically evaluate the complete data analytics workflow from data ingestion to 5
insight communication, using the CPD case study as reference.
Unit 5 – Statistical Analysis and Hypothesis Testing
Question Marks
Analyze the concept of hypothesis testing and examine its relevance in drawing 1
statistically valid conclusions from data.
Critically analyze the t-test mechanism and evaluate its applicability with a suitable 5
real-world example.
How would you compute the mean and standard deviation of Sales data? What do 3
these measures reveal about the distribution?
Evaluate the correlation coefficient between Advertising_Spend and Sales. 5
Interpret the nature and direction of the relationship.
Examine the core principles of effective data visualization and assess how they 5
enhance data-driven decision-making.
What is a null hypothesis (H0) and an alternative hypothesis (H1)? Provide an 1
example from a sales dataset context.
Explain Type I and Type II errors in hypothesis testing. What is the significance 3
level α?
What is a p-value? How do you use it to accept or reject the null hypothesis? 1
Distinguish between Pearson and Spearman correlation coefficients. When should 3
each be used?
Describe the Chi-square test. In what scenarios is it more appropriate than a t- 3
test?
What is skewness and kurtosis? How do they help characterize the shape of a 1
data distribution?
Explain the Central Limit Theorem. Why is it foundational for statistical inference in 3
analytics?
Describe multivariate analysis. How does multiple linear regression extend simple 3
linear regression?
Question Marks
What is confidence interval estimation? Construct a 95% confidence interval for a 3
given sample mean and standard deviation.
Analyze the residual plot in regression analysis. What patterns indicate that the 3
assumptions of regression are violated?
Critically evaluate how to interpret a correlation matrix for the Employee 5
Performance Dataset (EPD). Which variables are most significant?
Sample datasets used in question bank
Student Performance Dataset
Roll No Name Math Science English Section
101 Amit 78 85 82 A
102 Neha 65 70 72 B
103 Rahul 90 88 91 A
104 Sneha 55 60 58 B
105 Arjun 88 92 85 A
Employee Salary Dataset
EmpID Age Salary Department Experience
1 25 30000 HR 2
2 28 NaN IT 3
3 35 120000 IT 10
4 40 500000 Sales 15
5 29 32000 HR NaN
6 35 120000 IT 10
Sales Dataset
Month Sales Advertising_Spend Region
Jan 200 50 North
Feb 220 55 South
Mar 250 60 North
Apr 300 70 South
May 280 65 North
Customer Purchase Dataset ·
CustomerID Age Gender Purchase_Amount Membership
C1 25 M 2000 Gold
C2 30 F 3500 Silver
C3 22 F 1500 Gold
C4 40 M 5000 Platinum
C5 35 F 4000 Silver
Employee Performance Dataset
EmpID Experience Performance_Score Salary Department
1 2 60 30000 HR
2 5 75 50000 IT
3 7 85 70000 IT
4 3 65 40000 Sales
5 10 90 90000 IT