--- title: "Data warehouse and data mining" subject: "Information System" chapter: 4 key: "semester_8/information_system:data-warehouse-data-mining" source: https://www.bctengineer.com/q/semester_8/information_system/data-warehouse-data-mining site: BCT Engineering --- # Data warehouse and data mining Information System · Chapter 4: Decision support and Intelligent systems - Asked in: 2072 Asoj, 5 marks; 2071 Bhadra, 4 marks (asked 2 times) - Video: [Data Warehouse and Data Mining, Information management](https://www.youtube.com/watch?v=6_QLB0q_-vA) — DWIVEDI GUIDANCE - Answer checked by Claude Opus 5.5 (high effort) on 2026-10-14 - Page: https://www.bctengineer.com/q/semester_8/information_system/data-warehouse-data-mining - Whole chapter: https://www.bctengineer.com/md/semester_8/information_system/chapter-4 ## Question What is a Data Warehouse? Why do you require data warehouse along with operational databases? How is it related with data mining? (Discuss with a suitable diagram.) ## Short answer A data warehouse is an integrated, subject-oriented, time-variant, non-volatile repository for analysis, fed by ETL from operational systems. It avoids slowing operational DBs, consolidates historical data, and provides clean input for data mining, which discovers patterns like associations and clusters. ## Answer #### Data warehouse A **data warehouse** is a central repository of integrated, subject-oriented, time-variant and non-volatile data collected from many sources, kept for analysis and decision support (Bill Inmon's definition). - **Subject-oriented** - organised by subjects such as customer, sales. - **Integrated** - consistent formats from many sources. - **Time-variant** - stores historical data. - **Non-volatile** - data is not changed after loading. ```text Operational DBs --\ Files/ERP/CRM ----- ETL --> Data Warehouse --> OLAP, reports External data ----/ | Data mining ``` #### Why a data warehouse is needed with operational databases 1. Operational databases are tuned for fast transactions; heavy analytical queries would slow them. 2. They hold only current data, while analysis needs years of history. 3. Data is scattered over many incompatible systems; the warehouse integrates and cleans it. 4. Operational data is detailed and normalised; the warehouse is designed (star schema) for fast aggregation. 5. It gives a single consistent version of data for management and avoids interference with live operations. #### Relation with data mining **Data mining** discovers hidden patterns, trends and relationships (classification, clustering, association rules) in large data sets. The warehouse supplies clean, integrated, historical data, which is the ideal input for mining. Mining results (customer segments, fraud patterns, market baskets) are used in decisions and may be stored back in the warehouse. ```text Warehouse -> Select/Preprocess -> Mining -> Patterns -> Decisions ``` ## Check yourself The answer's key points as multiple-choice questions, with the correct option and why. 1. What does non-volatile mean for a data warehouse? A. Data is not changed after loading B. Data is updated every second C. Data is deleted after one day D. Data is stored only for one subject Answer: A. Non-volatile means loaded data is stable and not modified. (Key point: Characteristics) 2. What does time-variant mean? A. Only current data is held B. The warehouse stores historical data C. Data changes format daily D. Data is hidden Answer: B. Time-variant means history is kept for analysis. (Key point: Time-variant) 3. Why is a data warehouse needed alongside operational databases? A. Operational databases already contain years of clean integrated history B. Warehouses are needed to process live transactions C. Heavy analytical queries would slow transaction-tuned operational databases, which also hold only current data D. Operational databases are designed for aggregation Answer: C. Operational systems are tuned for fast transactions and hold current data; the warehouse is built for analysis. (Key point: Why separate) 4. How is data mining related to the warehouse? A. Mining replaces the warehouse B. Mining only uses raw unintegrated data C. They have no relationship D. The warehouse supplies clean, integrated, historical data as input for mining Answer: D. Mining runs on warehouse data and results can be stored back. (Key point: Relation to mining) 5. What does the ETL step do in the diagram? A. Moves data from operational sources into the warehouse after extracting, transforming and loading B. Deletes the sources C. Draws charts D. Mines patterns Answer: A. ETL feeds the warehouse from many sources. (Key point: ETL) ## Related in Information System - [Data mining and business knowledge](https://www.bctengineer.com/question.md?key=semester_8%2Finformation_system%3Adata-mining-business-knowledge) — Information System, chapter 4 - [Short note: data mining and warehousing](https://www.bctengineer.com/question.md?key=semester_8%2Finformation_system%3Adata-mining-and-warehousing-note) — Information System, chapter 4 - [Web mining with structure mining](https://www.bctengineer.com/question.md?key=semester_8%2Finformation_system%3Aweb-mining-structure-mining) — Information System, chapter 7 - [IS areas using AI / data analytics](https://www.bctengineer.com/question.md?key=semester_8%2Finformation_system%3Aai-areas-data-analytics) — Information System, chapter 4, asked 2× ## Related in other subjects - [Data warehouse architecture and components](https://www.bctengineer.com/question.md?key=semester_7%2Fdata_mining%3Adw-architecture-components) — Data Mining, chapter 1 - [Data warehouse schemas](https://www.bctengineer.com/question.md?key=semester_7%2Fdata_mining%3Adw-schema-types) — Data Mining, chapter 1 - [Data warehouse, data mart and need](https://www.bctengineer.com/question.md?key=semester_7%2Fdata_mining%3Awarehouse-datamart-need) — Data Mining, chapter 1 - [Short note: Noise in data warehouse](https://www.bctengineer.com/question.md?key=semester_7%2Fdata_mining%3Anoise-in-dw-note) — Data Mining, chapter 2