infn gathering

8/3/2019 Infn Gathering

1/42

Data Collection and Analysis

- Fact Finding

- Requirements Capture -

Techniques forInformation Systems Analysis


2/42

Requirements Discovery

Sample Existing Documentation

Observation

Questionnaires

Interviewing

Sampling

Overview


3/42


4/42

Results of Poor Understanding

The system may cost more than projected.

be delivered later than promised.

not meet the users expectations: so they not to use it.

be very costly to maintain and enhance.

be unreliable and prone to errors and downtime.

The reputation of the IT staff is tarnished

any failure, regardless of who is at fault, will be perceivedas a mistake by the team.


5/42

The Process of Requirements

Discovery

Problem discovery and analysis

Needs to be identified before a project begins.

Requirements discovery

Documenting and analyzing requirements

Requirements management


6/42

Requirements Discovery Fact-finding: the formal process of using:

research,

meetings,

interviews,

questionnaires,

sampling,

to collect information about

system problems, requirements, and

preferences.

It is also called information gatheringordata collection.


7/42

Fact-Finding Ethics

Fact-Finding often brings systems analysts into contact withsensitive information. E.g. Company plans

Employee salaries or medical history

Customer credit card, social security, or other information

Ethical behaviour means: Systems analysts must not misuse that information.

Systems analysts must protect that information from people whowould misuse it.

Otherwise: Systems analyst loses respect, credibility, and confidence of usersand management, impairing ability to do job

Organisation and systems analyst could have legal liability

Systems analyst could lose job


8/42

Four Fact-Finding Methods

Sampling of existing documentation.

Observation of the work environment.

Questionnaires. Interviews.


9/42


Quantitative

Data sources

Decision makingreports

Performance reports

Records

Data capture forms

Qualitative

Data sources

Memos Notices

Procedure manuals

Policy handbooks


10/42


The process of collecting a representativesample of documents, forms, and records. E.g. Organisation chart.

Memos and other documents that describe the problem Standard operating procedures for current system Completed forms Manual and computerised screens and reports Samples of databases

Flowcharts other system documentation


11/42

Sample Completed Forms

Can determine the type of data going into each field

Can determine the size of data field

Can determine which fields are not used or not always

used, Can see data relationships


12/42

Observation

a technique where the systems analyst either participates in, or watches a person

perform activities to learn about the system.

Advantages? Disadvantages?

Variation: work sampling - a technique thatinvolves a large number of observations taken atrandom intervals.


13/42

Need to decide which activities to observe

when to observe

whether to use Time or Event sampling whether to be a neutral observer or an active

participant

which recording techniques to use

Observation


14/42

Observation Guidelines

Determine the who, what, where, when,why, and how of the observation.

Obtain permission from appropriate

supervisors or managers.

Inform those who will be observed of the

purpose of the observation. Keep a low profile (dont intrude or

disrupt the work).


15/42

Observation Guidelines

Take notes during or immediatelyfollowing the observation.

Review observation notes with

appropriate individuals.

Don't interrupt the individuals at work.

Don't focus heavily on trivial activities. Don't make assumptions.


16/42

Questionnaires

Enable the analyst to collect information and

opinions from respondents.

Advantages?

Disadvantages?


17/42

Question types

Free-format (open) questions

A question is asked, and the respondent records their

personal answer in the space provided after the

question.

Fixed-format (closed) question

questionnaire containing questions that require

selecting an answer from predefined available

responses.


18/42

Fixed-Format (Closed) Questions

Multiple-choice questions Rating questions

Ranking questions

Rank the following transactions according to the amount of time you spend

processing them.

___ % new customer orders

___ % order cancellations

___ % order modifications

___ % payments

The implementation of quality discounts would cause

an increase in customer orders.

Strongly agree

AgreeNo opinion

Disagree

Strongly disagreeIs the current accounts receivable

report that you receive useful?

Yes

No


19/42

Fixed-Format (Closed) Questions

Checklists:simple, limitedresponses are permitted

Multipoint Ratings:use a linearscale to give responses on agradual basis.

Likert Scale: multipoint ratingswhere strength of agreement isidentified against a clearstatement

Semantic Differential Scale: placeopposites (such as easy-difficult)at endpoints - responses are ratedon a scale between the two.

Ranked Ordered Scale. Place a listof items in order against oneanother


20/42


21/42


22/42

Interviews

The systems analysts collect information fromindividuals through face-to-face interaction.

Can be used

For fact gathering: Find, Verify, Clarify facts

To generate enthusiasm

To get the end-user involved

To identify requirements To gather ideas and opinions

Advantages? Disadvantages?

The personal interview isgenerally recognised as the

most important and most often

used fact-finding technique.


23/42

Types of Interviews

Unstructured interview conducted with only a general goal or subject in

mind and with few, if any, specific questions.

The interviewer counts on the interviewee toprovide a framework and direct theconversation.

Structured interview

Where the interviewer has a specific set ofquestions to ask of the interviewee.


24/42

Interview question formation similar to questionnaires

Interviewing steps

Read background material and establish interviewing

objectives

Decide who to interview

Prepare the interview

Carry out the interview Prepare a written report

Review the report with the interviewee

Interviewing


25/42


26/42


27/42

Interview Questions

Open-ended v. closed-ended question One that allows the interviewee to respond in

any way that seems appropriate.

One that restricts answers to either specificchoices or short, direct responses.

Types of Questions to Avoid

Loaded questions Leading questions

Biased questions


28/42

Interview Questions

Interview Question Guidelines Use clear and concise language.

Dont include your opinion as part of the question.

Avoid long or complex questions.

Avoid threatening questions.

Dont use you when you mean a group of people.


29/42


30/42

Communicating With the User

Guidelines for CommunicatingApproach the Session with a Positive Attitude

Set the Other Person at Ease

Let Them Know You Are ListeningAsk Questions

Dont Assume Anything

Take NotesTo hear is to recognize that someone is speaking, to listen is to understand

what the speaker wants to communicate. (Gildersleeve 1978)


31/42

Body Language and Proxemics

Body language the nonverbal information wecommunicate.

Facial disclosure

Eye contact

Posture

Proxemics the relationship between people

and the space around them.

Intimate zonecloser than 0.5m Personal zonefrom 0.5-1m

Social zonefrom 1-4m

Public zonebeyond 4m


32/42

A Fact-Finding Strategy

1. Learn from existing documents, forms, reports, andfiles.

2. If appropriate, observe the system in action.

3. Given all the facts that already collected, design anddistribute questionnaires to clear up things that arent

fully understood.

4. Conduct interviews (or group work sessions).

5. (Optional). Build discovery prototypes for anyfunctional requirements that are not understood or for

requirements that need to be validated.

6. Follow up to verify facts.


33/42

to Consider In Planning Data Collection andAnalysis

What data should be gathered ?

Information analysed should be

Sufficient and

representative (of both standard cases and

exceptions)

In most cases you can only gather samples of datawhichever technique(s) you use.

General Issues

Th C t i t f th S l


34/42

Are you looking for statistical significance?

If so

Determine the data to be collected or described Determine the population to be sampled

Choose the type of sample

Decide on the sample size

The Constraints of the Sample

Used


35/42

Sample Size

How to Determine Sample Size: The size of the sample depends on how representative

you want the sample to be.

One simple and reliable formula for determining

sample size is

Sample size (SS) = 0.25 x (Certainty factor/Acceptable

error)

2


36/42

Sample Size

Sample size (SS) = 0.25 x (Certaintyfactor/Acceptable error)

2

The certainty factor depends on how certain you want to

be that the data sampled will not include variations not

in the sample.

The certainty factor is calculated from tables (available

in many industrial engineering texts). A partial example

is given here.

Desired Certainty Certainty Factor

95% 1.960

90% 1.645

80% 1.281


37/42

Example for Sampling

How to Determine Sample Size:

Suppose you want 90% certainty that a

sample of invoices will contain nounsampled variations.

SS = 0.25(1.645/0.10)2 = 68 (to zero dec. places)

So we need to sample 68 invoices to

get the desired accuracy.

E l f S li


38/42

Example for Sampling

Suppose we know from experience that one

in every ten invoices varies from the norm.

Based on this knowledge we can alter the above

formula by replacing the heuristic 0.25 with p(1-

p)

Wherep is the proportion of invoices with variances.

SS =p(1-p) (1.645/0.10)2.

SS = 0.10(1-0.10) (1.645/0.10)2 = 25 (to 0 dec places)

So we only need to sample 25 invoices to get the desired

accuracy


39/42

Convenience sampling Not based on probability and is unrestricted

Purposive sampling Based on specific criteria based on probability selected by the

analyst

Simple random selection Has no selection criteria

Complex random Data is probabilistic and is selected according to specific criteria

Systematic

Stratified

Clustering

Types of Sampling

S S S


40/42

Selecting the Sample Set

Consider two of these commonly used samplingtechniques: randomisation and stratification.

Randomisation:

has no predetermined pattern or plan for selecting sample

data.

In this example we therefore just randomly choose 25

invoices (regardless of the volume of the full set).

Stratification:

systematic sampling

attempts to reduce the variance of the estimates by spreading

out the sampling

for example, choosing documents or records by formula

S l Q lit ti D t


41/42

Sample Qualitative Data

There is no magic formula to help thesystems analyst set the sample size forinterviewing (for example).

The overriding variable that determines howmany people the systems analyst shouldinterview in depth is the timean interviewtakes.

A true in-depth interview and follow-upinterview is very time-consuming for both theinterviewer and participant.

S l Q lit ti D t


42/42

Sample Qualitative Data

A heuristic is:interview at least three people on every level

of the organisation,

and at least one from each of theorganisation's functional areas.

You do not have to interview more peoplejust because it is a larger organisation.

If the stratified sample is done properly, asmall number of people will adequatelyrepresent the entire organisation.

infn gathering

Documents