MauryaHub PYQ Practice

Tools in Data Science · End Term · Jan 2026 · AN

Tools in Data Science · End Term · Jan 2026 AN

← Course papers · Start practice / exam

Questions and published explanations below are available without starting a test. Some questions may not have a published solution yet.

0 / 27 published solutions

Question 2 MCQ · 1.0 marks

You construct a
Source diagram or notation
file. Which formatting rule must be followed to avoid parsing errors? Choose the correct answer.
A.
Use curly braces
Source diagram or notation
to define arrays.
B.
Use spaces for indentation, not tab characters.
C.
Wrap bash scripts in HTML
Source diagram or notation
tags.
D.
End every line with a semicolon
Source diagram or notation
.

A published solution is not available for this question yet.

Report a correction for Q2

Question 3 MCQ · 1.0 marks

A student types
Source diagram or notation
in their Bash terminal. If
Source diagram or notation
doesn't exist, what immediately occurs? Choose the correct answer.
A.
Shell throws a "File Not Found" error.
B.
SQLite creates a new empty database file
Source diagram or notation
.
C.
Interactive prompt asks for a database schema.
D.
Kernel triggers a script to repair the system path.

A published solution is not available for this question yet.

Report a correction for Q3

Question 4 MCQ · 1.0 marks

A student queries a remote 5GB Parquet file versus 5GB CSV. Parquet queries are much faster. Why? Choose the correct answer.
A.
Parquet files execute logic server-side before returning data.
B.
CSV files include encryption that DuckDB struggles to decode.
C.
Parquet uses columnar storage—DuckDB downloads only requested columns.
D.
Parquet uses multithreading to unpack data in browser memory.

A published solution is not available for this question yet.

Report a correction for Q4

Question 5 MSQ · 1.0 marks

When committing code to a public repository, which files should be added to
Source diagram or notation
? Choose all correct options.
A.Source diagram or notation
B.Source diagram or notation
C.Source diagram or notation
D.Source diagram or notation

A published solution is not available for this question yet.

Report a correction for Q5

Question 6 MCQ · 1.0 marks

You run
Source diagram or notation
. What happens? Choose the correct answer.
A.
OS captures stdout and redirects to
Source diagram or notation
, replacing existing contents.
B.
Bash captures both stdout and stderr, merging them into
Source diagram or notation
.
C.
Script accepts
Source diagram or notation
as input payload.
D.
OS alerts the cloud environment that Python started.

A published solution is not available for this question yet.

Report a correction for Q6

Question 7 MCQ · 1.0 marks

You fetch data with
Source diagram or notation
. What's the difference between
Source diagram or notation
and
Source diagram or notation
? Choose the correct answer.
A.Source diagram or notation
parses bools;
Source diagram or notation
treats everything as strings.
B.Source diagram or notation
returns raw string;
Source diagram or notation
decodes to dictionary.
C.Source diagram or notation
executes JavaScript files remotely in Python.
D.
No difference—both return strings initially.

A published solution is not available for this question yet.

Report a correction for Q7

Question 8 MCQ · 1.0 marks

A Vercel server crashes for 5 minutes. A Cloudflare Worker performs the same API logic. What happens to the Worker? Choose the correct answer.
A.
Worker also crashes—both share cloud infrastructure.
B.
Worker continues serving—runs on 300+ edge locations; single failures don't cause outages.
C.
Cloudflare reroutes traffic to the crashed Vercel server.
D.
Worker pauses and queues requests until Vercel recovers.

A published solution is not available for this question yet.

Report a correction for Q8

Question 9 MCQ · 1.0 marks

Ngrok's free tier doesn't support custom subdomains. What's localtunnel's advantage? Choose the correct answer.
A.
Free and open-source with no signup needed to expose servers.
B.
Automatically deploys local servers to production cloud hosting platforms.
C.
Provides built-in authentication to block unauthorized tunnel access attempts.
D.
Encrypts all traffic whereas Ngrok sends data in plaintext.

A published solution is not available for this question yet.

Report a correction for Q9

Question 10 MCQ · 1.0 marks

A developer tests a sorting function with Hypothesis. Which represents a valid invariant for any sorting function? Choose the correct answer.
A.
Output always equals
Source diagram or notation
for any input.
B.
Output contains same elements as input; no element exceeds its successor.
C.
Executes in under 10 milliseconds for all inputs.
D.
Input list modified in-place without creating new list.

A published solution is not available for this question yet.

Report a correction for Q10

Question 11 MCQ · 2.0 marks

In DuckDB, you run
Source diagram or notation
. One row has
Source diagram or notation
latency. How is the average calculated? Choose the correct answer.
A.
Ignores
Source diagram or notation
; calculates using only valid rows.
B.
Coerces
Source diagram or notation
to
Source diagram or notation
and includes it in sum.
C.
Throws runtime error demanding
Source diagram or notation
filter.
D.
Final aggregate result becomes
Source diagram or notation
.

A published solution is not available for this question yet.

Report a correction for Q11

Question 12 MCQ · 2.0 marks

A spatial join of 100K customers to 50 warehouses takes 45 minutes without index, 8 seconds with index. Why? Choose the correct answer.
A.
Spatial indexes compress geometries, reducing transfer time.
B.
Indexes use tree structures, avoiding testing every customer-warehouse pair.
C.
Indexes parallelize queries across multiple CPU cores.
D.
Indexes cache previous query results in memory.

A published solution is not available for this question yet.

Report a correction for Q12

Question 13 MCQ · 2.0 marks

Your UI on
Source diagram or notation
fetches from
Source diagram or notation
. Browser throws
Source diagram or notation
. How to resolve? Choose the correct answer.
A.
Use WebSockets instead of HTTP.
B.
Encrypt frontend payload before dispatching.
C.
Add
Source diagram or notation
to backend response headers.
D.
Add
Source diagram or notation
to fetch call.

A published solution is not available for this question yet.

Report a correction for Q13

Question 14 MCQ · 2.0 marks

Why pin dependencies (e.g.,
Source diagram or notation
) instead of leaving them unpinned? Choose the correct answer.
A.
Ensures reproducible pipelines; prevents updates from breaking code.
B.
Downloads packages faster from cloud index.
C.
Unpinned versions fail to execute in Docker images.
D.
Pinned versions compress faster into virtual RAM.

A published solution is not available for this question yet.

Report a correction for Q14

Question 15 MCQ · 2.0 marks

A data analyst is computing the distance between two delivery points stored as latitude/longitude pairs (in degrees). They try three approaches: 1. Euclidean distance directly on the raw (lat, lon) degree values. 2. The **Haversine** formula with Earth radius R = 6371 km. 3. Converting the coordinates to a **projected CRS** (e.g., UTM) first, then using Euclidean distance in metres. Which statement is **correct**?
A.
All three methods give the same distance because lat/lon already act like regular metric units everywhere on Earth.
B.
Euclidean on raw degrees is the most accurate for far-apart points because it skips trigonometry and avoids numerical error.
C.
Haversine gives km along Earth's curve, UTM gives metres on a plane, and raw-degree euclidean fails because 1\(^{o}\) longitude varies.
D.
UTM gives equally accurate distances anywhere on Earth, no matter which UTM zone the points actually fall in.

A published solution is not available for this question yet.

Report a correction for Q15

Question 16 MCQ · 3.0 marks

Fraud model: 99.2% test accuracy, 52% production accuracy. Dataset chronological,
Source diagram or notation
, rolling stats computed before splitting. What caused collapse? Choose the correct answer.
A.
Model overfitted—common when accuracy exceeds 95%.
B.Source diagram or notation
created temporal leakage; rolling stats before split caused feature leakage.
C.
Different hardware causes numerical precision differences.
D.
Fraud patterns evolve; historical models always fail on new data.

A published solution is not available for this question yet.

Report a correction for Q16

Question 17 MCQ · 3.0 marks

A Vercel edge function processes logs; needs 40 seconds but hits
Source diagram or notation
at 10 seconds. Why? Choose the correct answer.
A.
JSON array logic crashed on duplicate row.
B.
Serverless edge enforces strict timeout limits.
C.
Cloud router dropped IP variable array mid-execution.
D.
Edge functions disable multithreading for loops >1000 iterations.

A published solution is not available for this question yet.

Report a correction for Q17

Question 18 MCQ · 1.0 marks

Read the following passage carefully: A student builds a FastAPI proxy backend to fetch movie data from a public API, pass the data to an LLM to generate a summary, and return the result to a React frontend. Initially, the student notices that every time they refresh the browser, the LLM takes 5 seconds to summarize the exact same movie, running up their API bill rapidly. To fix this, the student adds an explicit
Source diagram or notation
parameter directly to the response header. The application speeds up drastically. However, when the student tries to test the API locally by fetching from the React UI (
Source diagram or notation
) to the backend (
Source diagram or notation
), the browser abruptly rejects the network connection and throws an error indicating missing headers. After resolving the browser block, the student requires the LLM to return the summary strictly as a JSON object. Instead of hoping the LLM formats it correctly, they utilize the API's native "Structured Outputs" constraint parameter to force the mathematical generation of valid JSON architecture. During load testing, the backend evaluates 5 web requests concurrently using asynchronous execution (
Source diagram or notation
). However, the entire server process completely halts when one specific downstream node hangs randomly for 60 seconds without resolving. Finally, a malicious user inputs trick instructions via the "Movie Title" input field, attempting to cleanly override the LLM's primary instructions. Based on the above data, answer the given subquestions.
Why does the browser reject the connection when React (
Source diagram or notation
) fetches from FastAPI (
Source diagram or notation
)? Choose the correct answer.
A.
Backend lacks
Source diagram or notation
for cross-port traffic.
B.
Single-host computing lacks multiple TCP ports.
C.
Browsers block all
Source diagram or notation
testing traffic.
D.
Backend Python script deleted the network thread.

A published solution is not available for this question yet.

Report a correction for Q18

Question 19 MCQ · 1.0 marks

Read the following passage carefully: A student builds a FastAPI proxy backend to fetch movie data from a public API, pass the data to an LLM to generate a summary, and return the result to a React frontend. Initially, the student notices that every time they refresh the browser, the LLM takes 5 seconds to summarize the exact same movie, running up their API bill rapidly. To fix this, the student adds an explicit
Source diagram or notation
parameter directly to the response header. The application speeds up drastically. However, when the student tries to test the API locally by fetching from the React UI (
Source diagram or notation
) to the backend (
Source diagram or notation
), the browser abruptly rejects the network connection and throws an error indicating missing headers. After resolving the browser block, the student requires the LLM to return the summary strictly as a JSON object. Instead of hoping the LLM formats it correctly, they utilize the API's native "Structured Outputs" constraint parameter to force the mathematical generation of valid JSON architecture. During load testing, the backend evaluates 5 web requests concurrently using asynchronous execution (
Source diagram or notation
). However, the entire server process completely halts when one specific downstream node hangs randomly for 60 seconds without resolving. Finally, a malicious user inputs trick instructions via the "Movie Title" input field, attempting to cleanly override the LLM's primary instructions. Based on the above data, answer the given subquestions.
What does
Source diagram or notation
do? Choose the correct answer.
A.
Forces backend to pause 3,600ms before replying.
B.
Authorizes caches to store/serve data for 60 minutes.
C.
Limits response visibility to 1 user.
D.
Creates SQLite database on client to store data.

A published solution is not available for this question yet.

Report a correction for Q19

Question 20 MCQ · 1.0 marks

Read the following passage carefully: A student builds a FastAPI proxy backend to fetch movie data from a public API, pass the data to an LLM to generate a summary, and return the result to a React frontend. Initially, the student notices that every time they refresh the browser, the LLM takes 5 seconds to summarize the exact same movie, running up their API bill rapidly. To fix this, the student adds an explicit
Source diagram or notation
parameter directly to the response header. The application speeds up drastically. However, when the student tries to test the API locally by fetching from the React UI (
Source diagram or notation
) to the backend (
Source diagram or notation
), the browser abruptly rejects the network connection and throws an error indicating missing headers. After resolving the browser block, the student requires the LLM to return the summary strictly as a JSON object. Instead of hoping the LLM formats it correctly, they utilize the API's native "Structured Outputs" constraint parameter to force the mathematical generation of valid JSON architecture. During load testing, the backend evaluates 5 web requests concurrently using asynchronous execution (
Source diagram or notation
). However, the entire server process completely halts when one specific downstream node hangs randomly for 60 seconds without resolving. Finally, a malicious user inputs trick instructions via the "Movie Title" input field, attempting to cleanly override the LLM's primary instructions. Based on the above data, answer the given subquestions.
What guarantee does "Structured Outputs" provide for LLM JSON generation? Choose the correct answer.
A.
Encrypts JSON output within browser.
B.
Guarantees 100% factually accurate summaries.
C.
Restricts token generation to prevent JSON syntax violations.
D.
Increases API timeout by 3600 milliseconds.

A published solution is not available for this question yet.

Report a correction for Q20

Question 21 MCQ · 1.0 marks

Read the following passage carefully: A student builds a FastAPI proxy backend to fetch movie data from a public API, pass the data to an LLM to generate a summary, and return the result to a React frontend. Initially, the student notices that every time they refresh the browser, the LLM takes 5 seconds to summarize the exact same movie, running up their API bill rapidly. To fix this, the student adds an explicit
Source diagram or notation
parameter directly to the response header. The application speeds up drastically. However, when the student tries to test the API locally by fetching from the React UI (
Source diagram or notation
) to the backend (
Source diagram or notation
), the browser abruptly rejects the network connection and throws an error indicating missing headers. After resolving the browser block, the student requires the LLM to return the summary strictly as a JSON object. Instead of hoping the LLM formats it correctly, they utilize the API's native "Structured Outputs" constraint parameter to force the mathematical generation of valid JSON architecture. During load testing, the backend evaluates 5 web requests concurrently using asynchronous execution (
Source diagram or notation
). However, the entire server process completely halts when one specific downstream node hangs randomly for 60 seconds without resolving. Finally, a malicious user inputs trick instructions via the "Movie Title" input field, attempting to cleanly override the LLM's primary instructions. Based on the above data, answer the given subquestions.
In
Source diagram or notation
, 4 nodes resolve in 1s, 1 hangs for 60s. What happens? Choose the correct answer.
A.
Resolves after 1s, assigns
Source diagram or notation
to hanging thread.
B.
Waits 60s for all nodes before returning results.
C.
Python cuts the hanging thread from server.
D.
Hanging thread throws HTTP 404 to avoid timeout.

A published solution is not available for this question yet.

Report a correction for Q21

Question 22 MCQ · 2.0 marks

Read the following passage carefully: A student builds a FastAPI proxy backend to fetch movie data from a public API, pass the data to an LLM to generate a summary, and return the result to a React frontend. Initially, the student notices that every time they refresh the browser, the LLM takes 5 seconds to summarize the exact same movie, running up their API bill rapidly. To fix this, the student adds an explicit
Source diagram or notation
parameter directly to the response header. The application speeds up drastically. However, when the student tries to test the API locally by fetching from the React UI (
Source diagram or notation
) to the backend (
Source diagram or notation
), the browser abruptly rejects the network connection and throws an error indicating missing headers. After resolving the browser block, the student requires the LLM to return the summary strictly as a JSON object. Instead of hoping the LLM formats it correctly, they utilize the API's native "Structured Outputs" constraint parameter to force the mathematical generation of valid JSON architecture. During load testing, the backend evaluates 5 web requests concurrently using asynchronous execution (
Source diagram or notation
). However, the entire server process completely halts when one specific downstream node hangs randomly for 60 seconds without resolving. Finally, a malicious user inputs trick instructions via the "Movie Title" input field, attempting to cleanly override the LLM's primary instructions. Based on the above data, answer the given subquestions.
How to force the cache to fetch fresh data instead of serving cached results? Choose the correct answer.
A.
Append unique query parameter (e.g.,
Source diagram or notation
) to mutate cache key.
B.
Run
Source diagram or notation
on backend to invalidate headers.
C.
Delete the payload encryption algorithm locally.
D.
Refresh the browser tab twice rapidly.

A published solution is not available for this question yet.

Report a correction for Q22

Question 23 MCQ · 2.0 marks

Read the following passage carefully: A student builds a FastAPI proxy backend to fetch movie data from a public API, pass the data to an LLM to generate a summary, and return the result to a React frontend. Initially, the student notices that every time they refresh the browser, the LLM takes 5 seconds to summarize the exact same movie, running up their API bill rapidly. To fix this, the student adds an explicit
Source diagram or notation
parameter directly to the response header. The application speeds up drastically. However, when the student tries to test the API locally by fetching from the React UI (
Source diagram or notation
) to the backend (
Source diagram or notation
), the browser abruptly rejects the network connection and throws an error indicating missing headers. After resolving the browser block, the student requires the LLM to return the summary strictly as a JSON object. Instead of hoping the LLM formats it correctly, they utilize the API's native "Structured Outputs" constraint parameter to force the mathematical generation of valid JSON architecture. During load testing, the backend evaluates 5 web requests concurrently using asynchronous execution (
Source diagram or notation
). However, the entire server process completely halts when one specific downstream node hangs randomly for 60 seconds without resolving. Finally, a malicious user inputs trick instructions via the "Movie Title" input field, attempting to cleanly override the LLM's primary instructions. Based on the above data, answer the given subquestions.
How to defend LLM from malicious "Movie Title" input overriding instructions? Choose the correct answer.
A.
Append integer limit string inside proxy cache.
B.
Place core rules in isolated 'System Role' vs 'User Role' inputs.
C.
Limit React input to 20 alphanumeric characters.
D.
Encrypt movie string in database before querying API.

A published solution is not available for this question yet.

Report a correction for Q23

Question 24 SUBJECTIVE · 2.0 marks

HUGGINGFACE_DOCKER_LATENCY **Scenario Context:** A developer deploys a Machine Learning text-classification service hosted dynamically on Hugging Face Docker Spaces. During production monitoring, the developer realizes the service boots painfully slowly and frequently crashes abruptly midway through container initialization. Additionally, the developer attempts to map API logs natively into an internal
Source diagram or notation
SQLite file locally within the cloud environment but is unsure if this breaks data retention policies. **Provided Dockerfile:**
Source diagram or notation
**Provided Backend Structure (**
Source diagram or notation
**):**
Source diagram or notation
Please construct a technical analysis diagnosing the container logic sequence, explicitly addressing the given areas. Based on the above data, answer the given subquestions.
**Docker Layer Caching :** Identify the critical sequencing error within the provided Dockerfile. Explain structurally how this specific order ruins layer caching, leading to severely inefficient build pipelines upon every minor code update. **NOTE:** Your answer should not exceed 200 words.

A published solution is not available for this question yet.

Report a correction for Q24

Question 25 SUBJECTIVE · 2.0 marks

HUGGINGFACE_DOCKER_LATENCY **Scenario Context:** A developer deploys a Machine Learning text-classification service hosted dynamically on Hugging Face Docker Spaces. During production monitoring, the developer realizes the service boots painfully slowly and frequently crashes abruptly midway through container initialization. Additionally, the developer attempts to map API logs natively into an internal
Source diagram or notation
SQLite file locally within the cloud environment but is unsure if this breaks data retention policies. **Provided Dockerfile:**
Source diagram or notation
**Provided Backend Structure (**
Source diagram or notation
**):**
Source diagram or notation
Please construct a technical analysis diagnosing the container logic sequence, explicitly addressing the given areas. Based on the above data, answer the given subquestions.
**Model Cold Starts :** The Hugging Face pipeline initializes globally at application startup. Explain fundamentally why executing parameter weights fetching dynamically during runtime creates massive "Cold Start" execution latency. **NOTE:** Your answer should not exceed 200 words.

A published solution is not available for this question yet.

Report a correction for Q25

Question 26 MCQ · 1.0 marks

Read the passage carefully and answer the subquestions. A student builds a browser-based data dashboard to avoid backend server costs. They compile DuckDB and Python into WebAssembly using Pyodide, enabling SQL queries to run directly in the browser. The dashboard downloads a 1.5GB Parquet database locally. Initially, it works well—Parquet's columnar format allows querying specific columns without lag. However, during a demo, critical issues emerge. First, clicking "Run Complex Query" freezes the entire UI for 10 seconds—animations halt and buttons become unresponsive. Second, attempting to use Python's
Source diagram or notation
library (
Source diagram or notation
,
Source diagram or notation
) crashes the WASM engine with "module unavailable" errors. Third, loading multiple dashboards across Chrome tabs triggers "Out-Of-Memory" exceptions, forcing the browser to terminate pages. The developer decides to abandon the browser approach and migrate the application to a Docker container running on
Source diagram or notation
.
Why does clicking "Run Complex Query" freeze the UI, blocking all clicks and animations?
A.
SQLite queries routinely disable client side graphics architecture by design intentionally
B.
WASM executes on the browser main thread blocking all UI rendering events
C.
React prevents overlapping interactions using internal thread locks dynamically and safely
D.
The server explicitly overrides the mouse tracking functionality remotely during execution

A published solution is not available for this question yet.

Report a correction for Q26

Question 27 MCQ · 1.0 marks

Read the passage carefully and answer the subquestions. A student builds a browser-based data dashboard to avoid backend server costs. They compile DuckDB and Python into WebAssembly using Pyodide, enabling SQL queries to run directly in the browser. The dashboard downloads a 1.5GB Parquet database locally. Initially, it works well—Parquet's columnar format allows querying specific columns without lag. However, during a demo, critical issues emerge. First, clicking "Run Complex Query" freezes the entire UI for 10 seconds—animations halt and buttons become unresponsive. Second, attempting to use Python's
Source diagram or notation
library (
Source diagram or notation
,
Source diagram or notation
) crashes the WASM engine with "module unavailable" errors. Third, loading multiple dashboards across Chrome tabs triggers "Out-Of-Memory" exceptions, forcing the browser to terminate pages. The developer decides to abandon the browser approach and migrate the application to a Docker container running on
Source diagram or notation
.
Why does Python code like
Source diagram or notation
fail when executing inside Pyodide/WebAssembly in the browser?
A.
The libraries are deprecated and removed in newer Python 3 versions
B.
Pyodide struggles to parse string formats inside terminal commands system limitations
C.
Browsers operate as sandboxes isolated from hardware without native OS bindings
D.
WASM requires sudo privileges to execute system commands inside Javascript environments

A published solution is not available for this question yet.

Report a correction for Q27

Question 28 MCQ · 1.0 marks

Read the passage carefully and answer the subquestions. A student builds a browser-based data dashboard to avoid backend server costs. They compile DuckDB and Python into WebAssembly using Pyodide, enabling SQL queries to run directly in the browser. The dashboard downloads a 1.5GB Parquet database locally. Initially, it works well—Parquet's columnar format allows querying specific columns without lag. However, during a demo, critical issues emerge. First, clicking "Run Complex Query" freezes the entire UI for 10 seconds—animations halt and buttons become unresponsive. Second, attempting to use Python's
Source diagram or notation
library (
Source diagram or notation
,
Source diagram or notation
) crashes the WASM engine with "module unavailable" errors. Third, loading multiple dashboards across Chrome tabs triggers "Out-Of-Memory" exceptions, forcing the browser to terminate pages. The developer decides to abandon the browser approach and migrate the application to a Docker container running on
Source diagram or notation
.
After moving the application to Docker with
Source diagram or notation
in the Dockerfile, running
Source diagram or notation
fails to connect at
Source diagram or notation
. What is missing?
A.
EXPOSE is documentation only developer must map ports using flag like dash p explicitly
B.
Docker container changes IP address sequentially requiring VPN proxy configuration for proper connectivity
C.
Localhost implicitly bans Docker from creating network bridges effectively for security protection reasons
D.
Execution lacks explicit SSL cryptographic certificate bindings required for local Docker network connections

A published solution is not available for this question yet.

Report a correction for Q28