Sampling Techniques in Systems Analysis

Why Sampling?

• Sampling in systems analysis helps contain costs, as examining every document or person is too expensive
• Sampling speeds up data gathering because a smaller sample reduces collection and analysis time
• Sampling improves effectiveness by allowing analysts to delve deeper with fewer data points, gaining better insights
• Sampling reduces bias by capturing a broader perspective and avoiding one-sided views in a well-designed sample

Designing a Sample in Systems Analysis:

• Determine the specific information required from the sample data
• Define the population to be sampled, specifying who or what the data will be collected from
• Choose the appropriate type of sample, such as random, convenience, purposive, or stratified samples
• Decide on the sample size, determining how many elements (people, documents, etc.) to include

Analyzing Quantitative Documents:

• Quantitative documents include reports used for decision making

Types of reports

• Sales reports summarize sales amounts and types
• Production reports include recent costs, inventory, labor, and plant information
• Summary reports provide background information, spot exceptions, and offer strategic overviews
• Service reports include delivery of services and reports on employee performance and customer satisfaction
• Performance reports compare actual vs. intended performance, assessing the gap and its trend
• Records provide periodic business updates, inspected for errors, design improvements, transaction analysis, and computer simplification opportunities
• Data capture forms involve collecting examples, noting the form type, documenting intended distribution, and comparing it with actual distribution

Analyzing Qualitative Documents:

• Focuses on affective, emotional, and motivational aspects of HCI and interpersonal relationships

Qualitative documents examples

• Key or guiding metaphors inform analysis and design
• Examination of insiders vs. outsiders mentality
• Understanding what is considered good vs. evil provides insights
• Graphics, logos, and icons in common areas or web pages
• A sense of humor conveyed in communications or documents

Further types of qualitative documents

• Email messages
• Memos
• Signs/posters on bulletin boards
• Corporate websites (interactivity)
• Manuals
• Policy handbooks
• Observation provides insight into what organizational members actually do and relationships between decision-makers

Analyst's Playscript:

• It involves observing decision-makers' behavior and recording actions using action verbs
• Examples of actions: Talking, Sampling, Corresponding, Deciding

STROBE (Structured Observation of the Environment):

• It is used for observing the decision-maker's physical environment
• It confirms or negates the organizational narrative through observing surroundings

Seven Concrete Observable Elements of STROBE:

• Office location: Consider who occupies the corner office or if key decision-makers are on separate floors
• Desk placement: Assess if it encourages communication or demonstrates power
• Stationary equipment: Note if the decision maker gathers/stores information personally and the size of the storage
• Props: Identify presence of PC, smartphone, or tablet use
• External information sources: Determine if trade journals or web resources are used
• Office lighting and color: Consider if it is for detailed work or casual communication and if colors are warm/inviting
• Clothing worn: Note whether clothing shows authority, such as conservative suits, or if uniforms are required

Link between STROBE and Decision-Maker Characteristics:

• Gathers information Informally: Warm lighting and colors
• Seeks Extraorganizational Information: Trade journals
• Processes Data Personally: PC or tablet in office
• Stores Information Personally: Equipment present
• Exercises Power: Desk for power
• Exhibits Credibility: Authoritative clothing
• Shares Information: Office accessible

Text Analytics:

• It uses software to analyze unstructured qualitative data from various sources
• Sources could be include transcripts of interviews
• Written reports
• Customer communications via email, wikis, blogs, chat rooms and social networking sites

Unstructured, qualitative, or “soft” data are generated commonly through:

• Blogs
• Chat rooms
• Questionnaires with open-ended questions
• Online discussions on the web
• Social media exchanges

Text Analytics applications:

• Provide insights into customer opinions about the organization
• It identifies company values and actions
• It understands customer or vendor motivations

Applying STROBE:

• STROBE identifies physical elements in a decision-maker's environment

STROBE observable elements:

• Office location (corner office)
• Desk placement (communication/power)
• Stationary equipment (size, storage)
• Props (PC, smartphone, etc.)
• External information sources (journals, web)
• Lighting and color (ambiance, communication/work)
• Clothing (formality, authority)
• It confirms or negates narratives from interviews/questionnaires

PROTOTYPES:

• Prototypes types are system models created to test design ideas
• Patched-Up Prototype entails a working, full-featured, but inefficient system
• Nonoperational Model functions to test design/outputs
• First-of-a-Series refers to pilot prototype tested at one or two sites
• Selected Features includes only essential features to test core functions
• User interaction should be honest, with feedback, and suggestions for improvement

DFD (Data Flow Diagram):

• DFD graphically represents data movement within a system, inputs, processes, data stores, and outputs
• It's useful in understanding system structure and processes

Definition of DFD

• It is a visual way of showing system inputs and outputs
• It visually represents data processes
• It shows Storage locations (data stores)
• It visually describes Interaction with external entities

Logical DFD – Advantages:

• Logical DFDs focus on business functions
• They lack commitment to technical design
• They improve communication with users
• They help uncover redundancy

Logical DFD - Disadvantages:

• Logical DFDs may not show implementation details
• They are less useful for developers needing code-level clarity

Physical Model – Advantages:

• Physical models include implementation details
• They identify manual vs automated processes
• They prepare systems for actual programming

Physical Model – Disadvantages:

• Physical models are less flexible
• They are more dependent on specific technology

DFD Symbols:

• External Entity is a double square
• Process is a rounded rectangle
• Data Flow is an arrow
• Data Store is an open-ended rectangle

Common DFD Errors:

• Arrows in the wrong direction
• Unlabeled arrows or processes
• Data stores connected directly to external entities
• Missing or extra processes
• Unbalanced decomposition

Primitive Process:

• Refers to a process is not exploded further in the DFD hierarchy
• It performs one clear function and is the lowest level in the DFD structure

Context Level / DFD 0 / Child Diagrams / Physical Diagrams:

• Context Diagram is the highest level DFD, with one process labeled 0, showing all external entities and data flows
• DFD 0 (Level 0) expands the context diagram into major subsystems
• Child Diagram details each process from DFD 0
• Physical Diagram shows actual system implementation, including manual steps and hardware

Unbalanced Decomposition:

• This occurs when a child diagram has different inputs/outputs than the parent process
• Therefore, all data entering or leaving the parent must match the child diagram's flows

DFD Partition:

• Involves dividing DFDs into separate modules for implementation
• It is based on user groups, execution time, task similarity, data consistency, or security
• It is represented by dashed lines around grouped processes

Event Modeling and Use Cases:

• Event Modeling creates DFDs based on business events (e.g., customer login)
• Use Cases define user actions and system responses
• Each use case becomes a process in the DFD

CRUD Functions:

• This tracks operations across processes
• Create: Add new record
• Read: Retrieve/view record
• Update: Modify existing record
• Delete: Remove record

Base / Derived Elements:

• Base elements are entered directly by the user
• Derived elements are calculated by the system

Data Structures – Algebraic Notation:

• Used to define the makeup of data structures:
• = means "is composed of"
• + means "and"
• {} represents repeating groups
• [] represents either/or conditions
• () represents optional elements

Content of Data Store:

• ID and Name
• Alias
• File Type
• Format
• Data Structure reference
• Volume (avg & max)
• Primary/Secondary Keys
• Comments for extra info

Continuous Elements:

• Continuous elements have a range of valid values instead of fixed options
• Examples: GPA (0.0-4.0) and Age

Data Dictionary:

• It is a central repository for all project information
• Includes data elements, flows, stores, and processes
• It also contains definitions, lengths, types, sources, and destinations
• It validates DFDs, generates XML, and creates screens/reports

Data Formats:

• It specifies how data appears or is stored
• Bit: 1 or 0
• Char, Varchar: Text
• Datetime: Date/time
• Currency: Money values
• Int, Smallint: Whole numbers
• Float, Real: Decimal numbers
• Autonumber: Auto-incrementing field

Data Element Characteristics:

• Includes ID & Name
• Aliases
• Description
• Type: Base or Derived
• Length
• Validation Criteria
• Default Values
• Formatting

Developing Data Stores:

• Define the data stores purpose
• Indicate whether storage is temporary/permanent
• Identify relevant reports/screens
• Use data dictionary structures
• Document growth, access method, and update frequency

Structured English:

• It is defined as using pseudo-code for sequence, decision, case, and iteration
• It clarifies logic and is readable by users

Decision Tables:

• Decision Tables consist of four quadrants: conditions, condition alternatives, actions, and rules
• They identify impossible scenarios, contradictions, or redundancy

Decision Trees:

• Decision trees map visual paths clearly showing paths for decisions
• Suited when the sequence of actions matters or the order processes varies
• Easier to follow than tables in some cases