Introduction to Data Science and Data Management and Analysis
SWE · Final · Spring 26
Section ALL
Read the questions as text
Read off the PDF by AI, so it may have mistakes. The PDF is the original.
Daffodil International University
Faculty of Science & Information Technology
Department of Software Engineering
Final Examination, Spring 2026
Course Code: DS 331
Course Title: Introduction To Data Science and Data Management & Analysis
Sections & Teachers: ALL
Time: 2:00 Hrs
Marks: 40
Answer ALL Questions
[The figures in the right margin indicate the full marks and corresponding course outcomes. All portions of each question must be answered sequentially.]
1. A large online retail company collects data from many sources such as customer purchases, website visits, payment gateways, and delivery systems. In the beginning, the company used a process where data was collected from all sources, processed in a separate server to clean and organize it, and then stored in the main data warehouse for reporting. This system worked well when the data volume was low. As the company grew, the amount of daily data increased significantly. Millions of records were generated every day, and the separate processing server became slow. Because of this delay, business reports were not updated quickly, and managers could not make timely decisions. To improve performance, the company changed its system so that raw data was first stored directly into a cloud-based data warehouse. After loading the data, cleaning and transformation tasks were performed inside the warehouse using cloud computing resources. This reduced the delay and improved the system's ability to handle larger amounts of data. Later, the company wanted to provide live sales dashboards, instant fraud detection, and real-time product recommendations for customers. To achieve this, they introduced a new system where data was captured continuously from user activities, sent to a streaming platform, processed immediately, and stored in a real-time analytics system.
Based on the above scenario, interpret how the company's data architecture evolved from ETL to ELT and then to real-time data processing. Compare the advantages of each stage and discuss the main challenges of maintaining scalability and high performance in real-time data systems. | Mark: 10 | CLO2 Level-5
2. At Hogwarts School, Harry Potter is asked by Professor Snape to build a neural network model to classify magical mushrooms as poisonous or edible using 10 numerical features. Harry creates a model with one hidden layer, sigmoid activation, and an SGD optimizer. He experiments with learning rates of 0.1, 0.01, and 0.001. During training, he observes that with 0.1 the loss quickly becomes infinite, with 0.001 the loss decreases very slowly and stays around 0.69, and with 0.01 the training loss falls to 0.32 with 91% training accuracy. However, the validation accuracy remains only 71%, indicating weak performance on unseen data. Even after tuning several parameters, Harry cannot improve the model's validation performance and asks Professor Snape to explain the behavior of the model and suggest improvements.
a. Evaluate how the different learning rates influence the model's training behavior, and explain why the learning rate of 0.01 provides better convergence compared to the other two values. | Mark: [Not specified]
b. Based on the training and validation results, evaluate the model's generalization performance and recommend suitable changes to improve its effectiveness on unseen data. | Mark: 10 | CLO2 Level-5
3. You are a data analyst at an AI learning platform. The dataset contains the following attributes: student_id, course_id, enrollment_date, total_watch_time, quiz_attempts, forum_posts, course_status
The platform tracks detailed student engagement metrics to help instructors improve course quality.
A. The academic director wants to understand how much time students typically spend watching course content in each course.
Write an SQL query to assess the average total watch time for each distinct course | Mark: 6 | CLO2 Level-5
B. To evaluate course popularity, the marketing team needs to know which courses have the highest number of enrollments.
Write an SQL query which can rank, counts the number of students enrolled in each course and orders the results in descending order.
C. Instructors want to identify the students who have successfully completed their courses. Write an SQL query to justify the list details of students whose course status is 'Completed'.
4. You are a data analyst working for a power distribution company. The company has recorded the monthly electricity consumption (in megawatt-hours, MWh) for a residential area during the year 2025:
Month | Consumption (MWh) | Month | Consumption (MWh)
January | 420 | July | 520
February | 400 | August | 500
March | 380 | September | 470
April | 360 | October | 430
May | 390 | November | 410
June | 450 | December | 440
A. Analyze the overall trend of electricity consumption throughout the year.
B. Detect the seasonal patterns (high or low).
C. Identify the monthly percentage change in consumption.
D. Examine the moving average for the sales data with a window size of 4 months. | Mark: 14 | CLO3 Level-4