Part A: Foundations
Introduction to data management information systems
Key takeaways
- Data management collects, stores, organises and maintains data so it stays accurate, accessible and secure throughout its life.
- The DIKW pyramid climbs from data (raw facts) to information (data given context), knowledge (patterns understood) and wisdom (judgement on how to act).
- An MIS captures, processes, stores and retrieves data, reports to every management level and controls performance against plans.
- A DBMS is layered into physical, logical and view layers, and is built from a query processor, storage manager, transaction manager and data dictionary.
- Relational databases suit structured data needing strict consistency; document and key-value stores suit varying or widely scaled data; graph databases suit data where the relationships are the point.
This paper moves from using information systems to building and overseeing the database layer underneath them. This opening chapter fixes the core vocabulary, from raw data up to organisational wisdom, before the rest of the paper builds a real database system on it.
What data management is
Data management is the discipline of collecting, storing, organising and maintaining data so it remains accurate, accessible and secure throughout its life, the foundation every information system (which turns that data into usable information) is built on.
The DIKW pyramid
Functions of MIS
- Data capture: recording transactions and events as they occur
- Data processing: transforming raw data into usable information
- Storage and retrieval: keeping data accessible for current and future use
- Reporting: presenting information to support decisions at every management level (see the types of information system in the next chapter)
- Control: monitoring performance against plans and flagging deviations
Where databases are used
- Banking and finance: account records, transactions, mobile-money ledgers
- Retail and e-commerce: inventory, orders, customer records
- Healthcare: patient records, scheduling, billing
- Education: student records, results, enrolment (the very system this app itself is built on)
- Government: civil registration, tax records, land records
Types of databases
| Type | Model |
|---|---|
| Relational database | Data organised into tables (relations) linked by keys |
| Hierarchical database | Data organised in a tree, each record having one parent |
| Network database | Data organised as a graph, a record can have multiple parents |
| Object-oriented database | Data stored as objects, as in object-oriented programming (see the chapter on object-based databases) |
| NoSQL database | Non-relational, flexible-schema storage for large-scale or unstructured data (see the chapter on emerging issues) |
The choice between database types follows from the data and how it is used. Relational databases suit structured data with clear relationships and strict consistency, such as accounting records. Document and key-value databases suit data whose structure varies or that must scale across many servers, such as product catalogues and session data. Graph databases suit data where the relationships themselves are the point of the analysis, such as fraud rings or social networks.
DBMS architecture
A Database Management System (DBMS) is typically organised in layers: the physical layer (how data is actually stored on disk), the logical layer (how data is structured, tables, relationships), and the view layer (how different users see the data, often a restricted or reshaped subset of the full schema), insulating each layer from changes in the ones below it.
Components of a DBMS
Facilities of a DBMS
- Data definition: creating and modifying the schema (the data definition language)
- Data manipulation: inserting, updating, deleting and querying data (the data manipulation language)
- Data security: controlling who can access or change what
- Data integrity: enforcing rules that keep data accurate and consistent
- Concurrency control: letting multiple users work with the database safely at once
- Backup and recovery: protecting against data loss
Exam tip
The DIKW pyramid is a favourite short-answer question, be ready to classify a given example (a raw sensor reading, a monthly trend report, an experienced manager's decision) against the correct level, the test is always how much processing and judgement has been applied to the raw data.
Self-check
(a) Define data management, and state the four levels of the DIKW pyramid in order. (5 marks)
(b) State three functions of an MIS. (3 marks)
(c) Distinguish a relational database from a NoSQL database. (3 marks)
(d) List the four components of a DBMS. (4 marks)
Show worked answer
(a) Data management is the discipline of collecting, storing, organising and maintaining data so it remains accurate, accessible and secure throughout its life. The DIKW pyramid, in order: Data, Information, Knowledge, Wisdom.
(b) Any three of: data capture; data processing; storage and retrieval; reporting; control.
(c) A relational database organises data into tables (relations) linked by keys, with a fixed schema. A NoSQL database is non-relational, with flexible-schema storage suited to large-scale or unstructured data.
(d) Query processor, storage manager, transaction manager, and data dictionary/catalog.