Chapter 16 - Systems Development and Data Analytics Flashcards
Phases of a System Development Life Cycle
- Initiation, feasibility, and Planning: understand the situation to see if it is feasible to create a solution and formulating a plan.
The project initiation phase includes promptly informing managers and employees about the project, assembling the project team (possibly including systems analysts, programmers, accountants, and users), training selected personnel to improve necessary skills and enhance communication among team members, and establishing project controls (e.g., by implementing a project scheduling technique such as PERT)
- Requirement analysis and definition:
a. Formal proposal for a new system is proposed to the IT steering committee.
b. Feasibility study happens: It should consider the activity to be automated, the needs of the user, the type and cost of equipment required, and potential benefits. FEASIBILITY AND COST ARE ACCESSED, and developing the logical requirements of a proposed system
c. STEERING COMMITTEE GIVES GO AHEAD. - System conceptual design
a. Mapping: data flow and storage diagrams and flowcharts.
b. System analysts involved. - Building and development
a. Actual program code and DB structure are written
b. Hardware is acquired, and physical structure is assembled - Testing and quality control
a. Methods:
i. Static testing: examine’s program code and its associated documentation without running program.
ii. Dynamic testing: involves executing code with a given set of test cases.
iii. White-box testing: tests internal structure instead of front end.
iv. Black box testing: tests functionality without any knowledge of the source code, tests only at user (blackbox) level.
v. Graybox testing: knows both internal data structure and also performs test to user.
b. Levels of tests
i. Unit testing: workability of specific section of codes or multiple items/subcomponents.
ii. Integration testing: test interaction of different components/modules.
iii. System testing or end to end: tests a completely integrated system to verify that the system meets its requirements.
iv. Acceptance testing: is conducted to determine whether the system meets the org needs and is ready for release. - Acceptance, installation and implementation
a. 4 types of conversion
i. In parallel.
ii. Direct change over.
iii. Pilot coversion: one branch, department or division at a time.
iv. Phased conversion: one function at a time. Oracle and Adaptive
v. Additional: prototype conversion involves developing and putting into operation successively more refined versions of the system until sufficient information is obtained to produce a satisfactory design.
b. Training and documentation are criticial.
PROGRAM CHANGE CONTROL
- Starts always from users.
- Save a copy of the production to the sandbox.
- Changes performed in sandbox version.
- When ready, it is named object code or executable code.
- Testing is done but not against production data, a special set of test data must be provided.
- User accepts or not (then make more adustments).
- Programmers move program to a holding area (they shouldn’t release themselves).
- Program supervisor reviews the program, and operations personnel move to production (compensating control since programmers cannot put fraudulent code into production).
End User Computing (limitations)
End user computing
- Involves user-created or user acquired systems that are maintained and operated outside of traditional information controls (like the queries in DWH)
- Environmental control risks are greater (copywright violation or unauthorized copies etc).
- Unauthorized access is another concern (lacks physical controls).
- Not adequate backup and contingency planning.
- Data might not be uniform/consistent across EUC applications.
- To handle that:
- Find controls and their intended functions
- One approach is to take an organization-wide inventory of major EUC applications.
- Risk assessment
- Review control of applications.
- Find controls and their intended functions
BUSINESS PROCESS REENGINEERING
- Involves a complete rethinking of how business functions are performed to provide value to a customer (radical innovation and a disregard of current jobs, hierarchies and reporting relationships.
BUSINESS PROCESS DESIGN AND PARTICIPANTS ROLES
- End users are the drivers of a new or redesigned process.
-
Because IT pervades every aspect of operations in a modern organization, the IT STEERING COMMITTEE must study each request for a new process or either approve it or deny it.
- Usually CIO or head of systems development are members. Executive mgmt. of each division is also involved.
- The committee members have an understanding of the interactions of the organizations current structure. (like the PC2 project)
- Once a new process or system has been approved, a project team is is assembled.
Artificial Intelligence (Benefits)
- Helps on Robotic Process Automation
-
How does it improve efficiency and effectiveness:
- Automate data entry and analysis (pdf reader)
- And it learns while it does that.
- Reduce fraud
- Review entire population of documents.
- Strengthen expenditure disbursement policies
- Check vast amount of documents to see if they are compliant with policies.
- Automate data entry and analysis (pdf reader)
Cloud Computing Benefits
- Fast access to software, reduce need for investment in IT infra, and use of “pay as you go” service.
- Uses do not have to buy, maintain or update computer hardware or software (done by 3rd party a lot).
- Subscribe only to the services needed.
- Users can access software from anywhere.
- Blockchain
- It is a distributed ledger that is encrypted, public and shared widely for anyone to view. Data is not centralized.
- A block is the current part of a blockchain
- Used primarily to verify transactions within digital currencies
- Cryptocurrency is a digital designed to be a medium of exchange using encryption to secure transactions, control the creation of additional units and verify transfer.
- Owning a bitcoin means that the user owns a piece of information (block) within the blockchain ledger.
- The blockchain allows users to transfer funds without the need of a financial institution intermediating.
- Miners then come into play confirming each new block of transaction seeing if they are legitimate. They are then paid bitcoins for thjeir effort.
- Used primarily to verify transactions within digital currencies
CONSENSUS MECHANISM maintained and validated by multiple agents.
SMART CONTRACTS
Are computerized transaction protocols that execute the terms of a contract. A smart contract is a collection of code and data that is deployed using crypto. THEY SATISFY COMMON CONTRACTUAL CONDITIONS WITHOUT NEEDING LAWYERS.
Data Approach to Quantify or Understand Risk (Delpy Approach, Predictive Analysis, EDA)
Delpy approach - Get opinion from set of experts and then share back to them so they can give feedback again and it should converge to an optimal solution.
Predictive analysis – uses data with statistical algorithmus and machine learning to draw insights in consumer behavior. Check out suggestions for a customer.
EDA (exploratory data analysis) - INSIDE DATA MINING APPROACH – different from traditional analysis since it doesn’t start from a given assumption/hypothesis. Seeks to:
- Maximize insight into data set.
- Uncover an underlying structure.
- Identify important variables.
- Check for outliers.
- Test underlying assumptions
- Develop parsimonious models
- Determine optimal factors.
- USE OF HISTOGRAMS AND SCATTER PLOTS.
BIG DATA: Concept, 5 V’s and limitations
BIG DATA
>> Process of mining volume amount of data from all sorts (unstructured on the beginning ) to transform it from low density to high density data (which has value). Data is usually in text format.
>> 4 Vs:
- Volume: large amount of data captured over time. IoT internet of things help on this since it owrks like real-time sensors and have the ability to transfer data over a network without requiring human to human or computer interactin. Might require multiple servers.
- Variety: structure data like the ones maintained in SQL (DWH), unstructured data (noSQL)
- Velocity: speed at which data has to be analysed./
- Veracity: to what extent data is representative of the population.
>> THE 5TH V: value.
>> Limitations of big data:
- User level data are incomplete (data restricted to target group that visited the website): a lot of untapped potential.
- Knowing on why results are like they are is ticky (you need more technical people).
- Data can contain noise (useless informatin)
- User level data requires interpretation prior to use (text to graph helps visualization though).
BI
Business intelligence (BI) has all of the following characteristics except
A.Displaying information in graphical format.
B.Giving immediate information about an organization’s critical success factors.
C.Focusing on strategic objectives.
Answer (D) is correct.
D.Providing advice and answers to top management from a knowledge-based system.
BI serves the needs of top management for managerial control and strategic planning. BI focuses on strategic (long-range) objectives and gives immediate information about a firm’s critical success factors. BI is not a program for providing top management with advice and answers from a knowledge-based (expert) system.
Statistical Control Charts
vs
Histogram
vs
The fishbone diagram (also called a cause-and-effect diagram)
Statistical control charts are graphic aids for monitoring the status of any process subject to acceptable and unacceptable variations during repeated operations. The chart consists of three horizontal lines plotted on a horizontal time scale. The vertical scale represents the appropriate quantitative measure. The center line represents the average range or overall mean for the process being controlled. The other two lines are the upper control limit and the lower control limit. The processes are measured periodically, and the values are plotted on the chart. If the value falls within the control limits, no action is taken. If the value falls outside the limits, the process is considered out of control, and an investigation is made for possible corrective action. Another advantage of the chart is that it makes trends visible.
The histogram displays a continuous frequency distribution of the independent variable in the form of a bar graph.
Fish bone is a total quality management process improvement technique. The format identifies and organizes the analysis of causation and helps to identify possible interactions among causes.
High-low method for Regression
(Ymax - Ymin)/(Xmax-Xmin) = COEFFICIENT
Then calcualte fixed cost for the formula.
