Business Intelligence
CSE · Midterm · Fall 26
Batch 64
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 Computer Science & Engineering
Midterm Examination, Fall 2026
Course Code: CSE-438, Course Title: Business Intelligence
Level: 4 Term: 3 Batch: 64
Time: 01:30 Hrs Marks: 25
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) An e-commerce company uses Hadoop MapReduce to identify its most common customer complaints. The following complaint logs have been collected:
Log ID | Complaint Keywords
Log 1 | Refund Damage Refund
Log 2 | Delay Refund Damage Refund
Log 3 | Refund Delay Damage
Illustrate how MapReduce processes these logs to calculate the final Complaint keyword counts and explain how this information could support the company's decision-making. [5, CO1]
2. A retail chain is comparing sales from its North and South branches before deciding where to invest. Their reports use different formats and contain repeated records. Management has decided to preprocess both datasets before consolidating them into a single table. They also decided not to discard any potential outliers.
Branch North
Day | Mon | Tue | Wed | Thu | Fri | Sat | Sun
Value (thousand Tk.) | 21 | 22 | 23 | NULL | 23 | 24 | 45
Branch South
Day | Monday | Tuesday | Wednesday | Thursday | Friday | Saturday | SUN
Value (Tk.) | 23,500 | 24,000 | 24,800 | -5,000 | 25,500 | 26,000 | 31,000
a) Apply appropriate data preprocessing techniques and present the consolidated table. Use branch-wise median replacement for missing and invalid values. [4]
b) Apply suitable statistical modeling to find out the branch with the most stable performance. Identify the most representative typical sales value for the selected branch based on outlier analysis. [6, CO2]
3. a) A hotel group wants to develop a data warehouse to analyze bookings across its properties. Currently, booking details are stored in one table, causing repeated guest, room, and hotel information.
Hotel Booking Records (October 2026)
Booking ID | Date | Guest | Membership | Hotel ID | Hotel | City | Room No. | Room Type | Nights | Bill (Tk.)
B01 | Oct 01 | Aisha | Gold | H01 | Garden Inn | Dhaka | 101 | Deluxe | 2 | 12,000
B02 | Oct 01 | Rafi | Silver | H01 | Garden Inn | Dhaka | 502 | Regular | 1 | 4,000
B03 | Oct 02 | Nabila | Gold | H02 | Bay View | Khulna | 101 | Suite | 2 | 22,000
B04 | Oct 03 | Aisha | Gold | H02 | Bay View | Khulna | 101 | Suite | 1 | 11,000
B05 | Oct 03 | Tania | Silver | H02 | Bay View | Khulna | 102 | Regular | 2 | 9,000
B06 | Oct 04 | Rafi | Silver | H01 | Garden Inn | Dhaka | 101 | Deluxe | 3 | 18,000
Apply appropriate data modeling techniques to construct a Snowflake Schema from the given dataset, clearly indicating the primary keys (PK), foreign keys (FK), and relationships between tables. [5, CO2]
4. a) An e-commerce company stores millions of sales records containing Customer ID, Product, Region, and Revenue. Its BI dashboard frequently calculates total revenue across different regions, but the reports take a long time to load. Moreover, managers access the same revenue reports multiple times daily.
Analyze different data storage methods and determine the most suitable approach for improving the dashboard's query performance. What additional technique could be used to handle repeated analytical queries more efficiently? Justify your answer. [5, CO3]