Indian Institute of Technology Madras
BS Degree Programme · Examination & Study Paper
Tools in Data Science
End Term · 31 Aug 2025 (Forenoon)
Examination Instructions & Guidelines
- Solve on notebook/paper first before checking solutions.
- MCQ: Exactly 1 option is correct. Negative marking applies.
- MSQ: One or more options correct. Select all to score.
- Scientific calculator allowed for all calculations.
- Official answer key is sealed at the end of this sheet.
- Q1MCQ2 Marks·Choose one option
When using uv for Python package management, which command sequence correctly creates a new project with a specific Python version and installs dependencies?
- A

- B

- C

- D

- A
- Q2MCQ1 Mark·Choose one option
In Git workflow for collaborative data science projects, what is the most appropriate sequence for integrating feature changes?
- A

- B

- C

- D

- A
- Q3MCQ1 Mark·Choose one option
When designing error handling in Python data processing pipelines, which approach generally offers the most robust and maintainable solution for unexpected runtime errors?
- AUsing only generic try-except blocks
- BUsing try-except-finally with specific exception types and proper logging
- CUsing pre-emptive condition checks (if-else) without exception handling
- DAllowing all exceptions to propagate to the top level
- Q4MCQ1 Mark·Choose one option
A data scientist needs to identify unique treatment combinations in a clinical trial dataset with 50,000 rows. Which pandas approach would be most memory-efficient for large datasets?
- A

- B

- C

- D

- A
- Q5MCQ1 Mark·Choose one option
A developer debugging an API integration notices intermittent 429 status codes. Which Chrome DevTools approach would best help identify the rate limiting pattern?
- AConsole tab with error filtering
- BNetwork tab with timing analysis and request throttling simulation
- CApplication tab for storage inspection
- DPerformance tab for bottleneck analysis
- Q6MCQ1 Mark·Choose one option
For a geospatial analysis project involving buffer operations, spatial joins, and interactive mapping, which combination of libraries provides the most comprehensive solution?
- AMatplotlib + NumPy + Pandas
- BGeoPandas + Shapely + Folium
- CPlotly + Seaborn + SciPy
- DBokeh + NetworkX + PyProj
- Q7MCQ1 Mark·Choose one option
When cleaning a dataset with company names like "Microsoft Corp", "Microsoft Corporation", "MSFT", which OpenRefine technique would be most effective?
- ASimple find and replace operations
- BClustering algorithms with fingerprinting and n-gram comparison
- CRegular expression matching only
- DManual standardization
- Q8MCQ1 Mark·Choose one option

- A

- B

- C

- D

- A
- Q9MCQ2 Marks·Choose one option
When designing a SQLite database for time-series sensor data with millions of records, which indexing strategy optimizes both query performance and storage efficiency?
- APrimary key index only
- BComposite index on (timestamp, sensor_id) with partial indexes for active sensors
- CIndividual indexes on all columns
- DNo indexing to save storage space
- Q10MCQ2 Marks·Choose one option
A data analyst needs to identify customers with purchase patterns similar to high-value segments. Which pandas operation most efficiently computes customer similarity?
- A

- B

- C

- D

- A
- Q11MCQ1 Mark·Choose one option
You are working with a very large CSV file that is too big to fit into memory. Which pandas method is the best way to read and process the file without running out of memory?
- A

- B

- C

- D

- A
- Q12MCQ1 Mark·Choose one option
A developer optimizing database-heavy applications needs to identify query bottlenecks and connection issues. Which monitoring approach provides comprehensive insights?
- AApplication logs only
- BDatabase query profiling + connection pool metrics + application performance monitoring
- CNetwork traffic analysis only
- DMemory usage tracking exclusively
Read the passage below, then answer Q13–Q15
Scenario 1: AI-Powered Research Assistant Development
Dr. Kumar is developing an AI research assistant for environmental science students. The system needs to process academic papers, generate summaries, and answer domain specific questions while encouraging critical thinking.
Based on the above data, answer the given subquestions.
Q13MCQ1 Mark·Choose one optionThe research assistant uses a system prompt to guide its behavior. Which system prompt design principle is most important for educational effectiveness?
- AProviding direct answers to maximize efficiency
- BBalancing information delivery with Socratic questioning to promote learning
- CLimiting responses to prevent information overload
- DUsing technical jargon to maintain academic rigor
- Q14MSQ2 Marks·Choose all correct options
When students submit complex research queries, which factors most significantly impact the LLM's operational costs and response quality? (Select all that apply)
- AToken count of input prompts and generated responses
- BComplexity and specificity of the query requiring deeper reasoning
- CTime of day when queries are submitted
- DContext window utilization for multi-turn conversations
- EStudent's academic level
- Q15MCQ1 Mark·Choose one option
Which prompt engineering technique would generate the most pedagogically valuable response for environmental research?
- A"Explain climate change impacts"
- B"Write about environmental issues"
- C"Analyze the interconnected effects of ocean acidification on marine ecosystems, considering pH changes, species adaptation, and food web dynamics. What research methodologies would be most appropriate for studying these relationships?"
- D"Discuss environmental problems briefly"
Read the passage below, then answer Q16–Q20
Scenario 2: Smart City Traffic Management System
A metropolitan city is implementing a real-time traffic management system that processes data from IoT sensors, traffic cameras, and GPS devices to optimize traffic flow across 500+ intersections.
System Architecture:
● IoT sensors at intersections (Lat/Long coordinates provided)
● Central processing hub: (28.6139^o N, 77.2090^o E)
● Real-time data streams: traffic density, weather conditions, accident reports
● ML models for traffic prediction and route optimization
Based on the above data, answer the given subquestions.
Q16MCQ1 Mark·Choose one optionThe development team uses Git for version control of traffic algorithms. When a critical traffic routing bug is discovered in production, which Git workflow ensures rapid deployment of fixes while maintaining code integrity?
- A

- B

- C

- D

- A
- Q17MCQ2 Marks·Choose one option
Traffic engineers need to analyze historical data patterns. Which command-line approach efficiently processes large log files to extract peak traffic periods and route preferences?
- A

- B

- C

- D

- A
- Q18MCQ2 Marks·Choose one option
For deploying the traffic management system across distributed edge computing nodes at intersections, which containerization strategy optimizes resource utilization and ensures consistent performance?
- ASingle monolithic container with all services
- BMulti-stage Docker builds with service mesh architecture and resource constraints
- CVirtual machines for each intersection
- DDirect installation on edge hardware without containerization
- Q19MCQ1 Mark·Choose one option
When optimizing traffic routes for emergency vehicles during peak hours, which algorithmic approach best balances computational efficiency with real-time requirements?
- ABrute force calculation of all possible routes
- BA pathfinding with dynamic edge weights based on real-time traffic data
- CRandom route selection
- DStatic pre-computed emergency routes
- Q20MCQ1 Mark·Choose one option
The system calculates distances between intersections for route optimization. What is the primary computational advantage of implementing the Haversine formula in the traffic management context?
- AProvides accurate great-circle distances for GPS coordinates without requiring road network data
- BCalculates exact travel time including traffic conditions
- CDetermines elevation changes for fuel efficiency
- DMeasures road surface quality for vehicle routing
Read the passage below, then answer Q21–Q30
Scenario 3: E-commerce Platform Log Analysis
You're analyzing Apache access logs from a major e-commerce platform. The logs contain customer browsing patterns, purchase behaviors, and system performance metrics during Black Friday sales.
Log Format Specification:

Analysis Requirements:
● Filter purchases (POST to /checkout/) during peak hours (12:00-15:59)
● Identify mobile vs desktop traffic patterns
● Track successful transactions (status codes 200-299)
● Analyze product category performance
Based on the above data, answer the given subquestions.
Q21MCQ1 Mark·Choose one optionWhich of the following Apache log entries correctly represents a successful mobile checkout transaction that occurred during the peak analysis window (12:00–15:59)?
- A

- B

- C

- D

- A
- Q22MCQ1 Mark·Choose one option
To extract essential fields for the analysis "mobile checkout transactions between 12:00-15:59 with successful status codes", which data elements are required?
- AIP, UserAgent, Referrer, VirtualHost
- BTimestamp, HTTP_METHOD, URL, StatusCode, UserAgent
- CRemoteUser, AuthUser, ResponseSize, ServerIP
- DIP, Timestamp, Referrer, ResponseSize
- Q23MCQ1 Mark·Choose one option

- A%d/%b/%Y:%H:%M:%S
- B%d/%b/%Y:%H:%M:%S %z
- C%Y-%m-%d %H:%M:%S %Z
- D%d-%b-%Y:%H:%M:%S GMT
- Q24MCQ1 Mark·Choose one option
To identify checkout transactions occurring between 12:00 and 15:59, which Python condition correctly validates the time range?
- A

- B

- C

- D

- A
- Q25MCQ1 Mark·Choose one option

- A

- B

- C

- D

- A
- Q26MCQ1 Mark·Choose one option
To identify mobile traffic from the UserAgent string, which Python approach provides the most reliable mobile device detection?
- A

- B

- C

- D

- A
- Q27MCQ1 Mark·Choose one option
For determining successful transactions, which condition correctly identifies HTTP status codes indicating success?
- A

- B

- C

- D

- A
- Q28MCQ1 Mark·Choose one option

- A

- B

- C

- D

- A
- Q29MCQ1 Mark·Choose one option
If the transaction count appears unexpectedly high, which systematic debugging approach would identify the root cause?
- AAssume data corruption and discard results
- BValidate each filter condition independently: time range → HTTP method → URL pattern → status code
- CReduce the dataset size randomly
- DCheck only the timestamp filter
- Q30MCQ1 Mark·Choose one option
When processing very large Apache log files (10GB+), which of the following Python approaches offers the best memory efficiency?
- A

- B

- C

- D

- A
Read the passage below, then answer Q31–Q34
Scenario 4: Academic Research RAG System with Multi-Modal Analysis
A research university implements an advanced RAG system that processes academic papers, datasets, code repositories, and experimental results to support interdisciplinary research across STEM fields.
Advanced Multi-Modal RAG Architecture:
● Text Processing: Research papers, grants, technical documentation
● Code Analysis: GitHub repositories, computational notebooks, algorithm implementations ● Data Integration: Experimental datasets, simulation results, sensor data
● Visual Processing: Figures, charts, experimental images, technical diagrams
● Semantic Linking: Cross-reference relationships between concepts, methods, and findings Based on the above data, answer the given subquestions.
Q31MCQ1 Mark·Choose one optionWhat is the primary advantage of multi-modal RAG systems over text-only approaches in academic research contexts?
- AReduced computational complexity for information retrieval
- BComprehensive understanding through integration of textual concepts, visual data, and computational methods
- CSimplified system architecture and maintenance
- DLower storage requirements for research materials
- Q32MCQ1 Mark·Choose one option
Which system prompt design best supports interdisciplinary research collaboration and knowledge synthesis?
- A

- B

- C

- D

- A
- Q33MCQ1 Mark·Choose one option
A researcher asks: "How can machine learning techniques be applied to optimize renewable energy grid integration, and what interdisciplinary approaches show promise?" Which response demonstrates the most effective research synthesis approach?
- A"Machine learning can predict energy demand and optimize grid operations through various algorithms."
- B"Here are 10 recent papers on ML applications in renewable energy: [lists papers with summaries]"
- C"Let's explore the intersection of several fields: What specific grid integration challenges interest you? ML approaches range from demand forecasting (time series analysis) to real-time optimization (reinforcement learning), while power systems engineering provides domain constraints. Materials science advances in energy storage create new optimization opportunities. Which aspect aligns with your research focus?"
- D"This is a complex topic requiring extensive literature review across multiple disciplines."
- Q34MCQ1 Mark·Choose one option
Which processing sequence correctly describes the multi-modal RAG system's comprehensive research synthesis workflow?
- AResearch Query → Text Search Only → Generate Literature Summary
- BResearch Query → Keyword Matching → Return Most Recent Papers
- CResearch Query → Multi-Modal Embedding (text + code + visual + data) → Cross-Disciplinary Retrieval → Concept Mapping → Synthesized Research Insights with Method Integration
- DResearch Query → Database Search → Template Response Generation