Decision Support and Data Integration Systems

Course Description

Decision Support and Data Integration Systems covers current tools and techniques available to support managerial decision-making, with analysis and practice in the building and use of decision support systems and expert/knowledge-based systems.

Within the SCNS taxonomy, ISM is the Information Systems Management prefix. The University of West Florida publishes this at 3 semester hours through the Department of Business Administration, College of Business. It is offered at approximately 4 Florida institutions.

⚠ The statewide inventory calls this number "Decision Support Systems"; UWF adds "and Data Integration Systems," and the addition is the more current half. Decision support in practice now means getting data out of several systems that were never designed to talk to each other — and that integration problem is usually harder than the analysis it feeds.

This course sits at an interesting historical junction, and it is worth naming. "Decision support systems" and "expert systems" are terms from the 1970s and 1980s, when the ambition was to encode expert reasoning as explicit rules. That programme largely stalled, and the current generation of AI took a different route entirely — learning patterns from data rather than encoding rules from experts.

⚠ The older material is not obsolete, and the reason is worth understanding. Rule-based systems are auditable: you can say exactly why a decision was reached. Machine-learned models frequently cannot, and in regulated settings — credit, insurance, hiring, healthcare — explainability is a legal requirement rather than a preference. Knowing both approaches, and when each is appropriate, is genuinely current professional knowledge.

The practical core of the course is that decisions are made on data that lives in the wrong places. Every real organisation has an ERP that does not match the CRM, a finance system with a different customer identifier, and a critical spreadsheet on somebody's desktop. The work is reconciling them.

⚠ The contact-hour figure is derived — the University of West Florida publishes none

UWF's catalog publishes a credit value in semester hours, the college and department, and prerequisites. It does not publish contact hours, a lecture and laboratory split, or terms of offering for any course. Every contact-hour value in a UWF guide in this repository is derived. The figure applies the standard lecture convention of 15 contact hours per credit, giving 45 hours for 3 semester hours. ⚠ Note that ISM courses are project-heavy, and scheduled contact hours understate the real time commitment considerably — budget for project work outside class. Confirm the meeting schedule and delivery mode with the department.

Learning Outcomes

Required Outcomes

• Describe the decision-making process and the types of managerial decision.
• Distinguish structured, semi-structured, and unstructured decisions.
• Describe the components and architecture of a decision support system.
• Describe model-driven, data-driven, and knowledge-driven decision support.
• Build and use a decision support model for a business problem.
• Apply what-if, sensitivity, and scenario analysis to a decision model.
• Apply optimisation techniques to a constrained business problem.
• Describe expert and knowledge-based systems and their architecture.
• Describe knowledge representation and rule-based inference.
• Describe the limits of rule-based systems and why they stalled.
• Compare rule-based and machine-learned approaches and their trade-offs.
• Explain why explainability matters in regulated decision contexts.
• Describe data integration and the problems it addresses.
• Describe extract, transform, and load processes.
• Describe data warehouses, data marts, and data lakes.
• Describe dimensional modelling and star schemas.
• Describe master data management and identity resolution.
• Describe data quality dimensions and their assessment.
• Describe data governance roles and responsibilities.
• Build and evaluate an integrated view across multiple data sources.
• Describe group decision support and collaborative decision-making.
• Evaluate a decision support system in business terms.

Optional Outcomes

• Apply the material to a real organisation or a case study.
• Use AI-assisted tools appropriately and verify their output.
• Communicate technical findings to a non-technical audience.
• Work effectively in a project team with defined roles.
• Build a portfolio artefact suitable for showing an employer.

Major Topics

Required Topics

• The decision-making process
• Structured, semi-structured, unstructured decisions
• DSS components and architecture
• Model-, data-, and knowledge-driven support
• Building decision models
• What-if, sensitivity, and scenario analysis
• Optimisation under constraints
• Expert and knowledge-based systems
• Knowledge representation and inference
• Limits of rule-based systems
• Rule-based versus machine-learned approaches
• Explainability in regulated contexts
• Data integration problems
• Extract, transform, load
• Warehouses, marts, and lakes
• Dimensional modelling and star schemas
• Master data management and identity resolution
• Data quality dimensions
• Data governance
• Building an integrated view
• Group and collaborative decision support
• Evaluating a DSS

Optional Topics

• Applied case studies and live organisations
• AI-assisted tooling and its verification
• Communicating findings to non-technical stakeholders
• Team project practice
• Portfolio development

Resources & Tools

• The text the instructor assigns — Sharda, Delen and Turban's Analytics, Data Science and Artificial Intelligence: Systems for Decision Support is the direct descendant of the classic text for this course and covers both generations of the material.
• Excel Solver and the analysis add-ins — free with the UWF Microsoft 365 licence; the optimisation and what-if material is genuinely well served by Solver, which most students never discover.
• Power BI Desktop and Power Query — free; Power Query is the most accessible real ETL tool available to a student, and the integration concepts become concrete immediately.
• Kimball Group's dimensional modelling resources — free articles; the star schema material in almost every textbook descends from Kimball.
• SQLite or PostgreSQL — free; joining data from several sources is best learned by actually doing it.
• Data.gov, Florida open data, and Kaggle — free; multiple real datasets that genuinely do not line up, which is the whole point.
• The UWF library's business databases — included in enrolment; practitioner sources such as Gartner and industry reports are behind paywalls a student already has, and most never find out.
• LinkedIn Learning — frequently free through UWF or a Florida public library card; strong on specific tools, weak on judgement. Use it for the software, not the thinking.

Career Pathways

• Business or systems analyst (SOC 15-1211, computer systems analyst) — the most common destination for this major, and the role this prefix is most directly built for: translating business needs into system requirements.
• Data analyst and business intelligence analyst (SOC 15-2051, data scientist; SOC 13-1111, management analyst) — the fastest-growing destination, and the reason the analytics courses in this prefix matter.
• Information security analyst (SOC 15-1212) — strong demand and strong pay in Florida, particularly around defence contracting in the Panhandle and Space Coast. Note that this prefix teaches security management, not penetration testing — a real and separate career.
• Database administrator and data engineer (SOC 15-1242, 15-1243).
• IT project manager — commonly entered after several years as an analyst; PMP or CAPM certification is the usual credential.
• ERP functional consultant — SAP, Oracle, Workday, Microsoft Dynamics; a well-paid path that is under-advertised to undergraduates.
• ⚠ Florida employer landscape: defence and aerospace contractors (Lockheed Martin, Northrop Grumman, L3Harris, and the Navy presence around Pensacola), healthcare systems (AdventHealth, Orlando Health, BayCare, Baptist, Ascension Sacred Heart), hospitality and theme parks (Disney, Universal, the cruise lines out of Miami, Port Canaveral and Tampa — all of which run substantial IT organisations), financial services (Raymond James in St. Petersburg, Fidelity in Jacksonville), logistics, and state and county government.
• ⚠ Security clearance is a genuine career asset in this state. Northwest Florida's defence concentration means clearable candidates have access to roles others do not — worth knowing early, because the process is slow and starts with an employer sponsoring you.

AI Integration

Information systems is one of the fields where generative AI has changed working practice fastest, and pretending otherwise would not serve a student. The honest position is that these tools are genuinely useful, genuinely unreliable, and that the professional remains accountable for the output regardless of what produced it.

• Where AI tools help in this domain: drafting and explaining code and SQL, generating test data, summarising documentation and standards, producing first-draft requirements and process descriptions, writing spreadsheet and BI formulas, and explaining an unfamiliar error message — which is a large share of what stalls a beginner.
• ⚠ Where they fail, specifically: they invent plausible functions, libraries, and API endpoints that do not exist; they produce code that runs and is subtly wrong; they are confidently wrong about anything niche, recent, or organisation-specific; and they cannot know your data, your business rules, or your regulatory constraints unless you tell them.
• ⚠⚠ The failure mode that matters most here is a wrong answer that looks right. A query that returns rows is not a query that returns correct rows. Verify against known values before you trust an analysis, and be especially careful with joins, filters, date handling, and anything involving aggregation.
• ⚠⚠⚠ Never paste confidential, personal, or regulated data into a public AI tool. Organisational data, customer records, health information, and anything under FERPA, HIPAA, GLBA or PCI-DSS must not leave a controlled environment. This is a genuine and common way that early-career employees cause serious incidents, and "I did not know" is not a defence that helps anyone. Use synthetic or anonymised data for anything you take outside a sanctioned system.
• Academic integrity: the course policy governs, and policies differ. Some instructors encourage AI use with disclosure, some restrict it to specified tasks, some prohibit it. Read the syllabus and ask if it is unclear — assuming is how students end up in an integrity process.
• ⚠ The professional argument for learning without it first is real, not moralising. If you cannot read and evaluate the output, you cannot catch it when it is wrong — and an employer is paying for the judgement, not the typing. Use AI to go faster on things you understand; do the learning yourself.
• Cite and disclose what you used, when the policy asks. This is becoming a normal professional expectation as well as an academic one.

Special Information

⚠⚠ ISM3011 is the gateway to the entire prefix

• Every undergraduate ISM course at UWF except ISM3011 itself requires it. ISM3116, ISM3323, ISM4113, ISM4320, ISM4321, ISM4400, ISM4481 and ISM4483 all name it. Nothing else in the prefix opens without it.
• ⚠ That makes it a single point of failure in your schedule. Delaying or failing ISM3011 delays everything downstream, and there is no alternative route around it. Take it as early as you are eligible.
• Only ISM3323 offers an alternative — it accepts ISM 3011 OR COP 2253, a programming route. It is the sole exception in the prefix.
• ⚠ ISM3011 itself has an unusual prerequisite: not a course, but completion of 45 hours of college coursework. See the note on standing requirements below.

⚠⚠ The statewide title and the UWF title differ — read the description, not the name

• Statewide inventory title: Decision Support Systems. UWF publishes it as: Decision Support and Data Integration Systems.
• This is an extension rather than a divergence. UWF adds data integration to the classical decision support content, which reflects how the field actually developed — the analysis was never the bottleneck; getting consistent data into one place was. Treated here as one subject.
• ⚠ The practical advice is the same in every case of title drift: carry a syllabus. A transfer evaluator matching on the number will accept it; one matching on the title may query it, and one comparing content may reach a different conclusion again.
• Search on the number, not the name, when looking for equivalents at other institutions. SCNS equivalency runs on the number and the content.

⚠⚠ Why "garbage in, garbage out" is the operative principle here

• A decision support system is only as good as the data feeding it, and integrated data is where quality problems become visible — the same customer appears three times with three spellings, dates arrive in two formats, and a "region" field means different things in two systems.
• ⚠ Identity resolution is the underrated hard problem. Deciding whether two records refer to the same person or company is genuinely difficult, consequential, and rarely taught as the discipline it is.
• Data quality has dimensions worth naming: accuracy, completeness, consistency, timeliness, validity and uniqueness. Measuring them is what turns a complaint into a fixable problem.
• ⚠⚠ Governance is the part that fails organisationally rather than technically. If nobody owns a data element, nobody fixes it — and naming an accountable owner is more effective than any tool.

⚠ Rule-based systems, machine learning, and when each is right

• Rule-based systems are transparent, auditable, and behave predictably — you can state exactly why an output occurred. They are brittle, expensive to maintain, and cannot handle situations their authors did not anticipate.
• Machine-learned models handle complexity and novelty far better and frequently cannot explain themselves, can encode bias present in their training data, and fail in ways that are hard to anticipate.
• ⚠⚠ In regulated decisions the choice may not be yours. Adverse credit decisions in the United States require specific reasons to be given; employment, insurance and healthcare decisions carry their own constraints. "The model said so" is not an adequate explanation, legally or ethically.
• Hybrid designs are common in practice — a model produces a score, rules enforce policy limits, and a human decides at the boundary. That shape is worth recognising, because it is what most deployed systems actually look like.

Certifications worth knowing about

• ⚠ A degree and a certification do different jobs. The degree is the durable credential; certifications are current, specific, and expire. Employers in this field ask for both, and neither substitutes for the other.
• Analytics and data: Microsoft Power BI Data Analyst (PL-300), Tableau Desktop Specialist, Google Data Analytics, AWS and Azure data certifications. Several have free or heavily discounted student pricing — ask the College of Business.
• Security: CompTIA Security+ is the standard entry credential and is frequently a hard requirement for defence-adjacent work under DoD 8570/8140; CISSP and CISM are management-level and require documented experience.
• Project and process: CAPM, PMP, and the Scrum credentials.
• ⚠ Do not collect certifications instead of building things. A portfolio of real projects — a dashboard, a database, an analysis with a written recommendation — outperforms a list of badges in almost every hiring conversation in this field.

Transfer, articulation, and how Florida course levels work

In the Florida Statewide Course Numbering System the first digit is the level: 1 and 2 are lower division, 3 and 4 upper division, 5 and above graduate. ⚠ A lower-division course generally cannot satisfy an upper-division requirement, which matters in this prefix — ISM2000 and ISM3011 both introduce information systems, and only the second is upper division.

⚠⚠ Business programmes add a layer that SCNS does not. Many Florida business colleges are AACSB accredited, and AACSB programmes commonly limit how much upper-division business coursework may transfer in — frequently requiring a substantial share to be taken in residence. A course can articulate under SCNS and still not count toward the major. Check the receiving programme's residency rule before you rely on a transfer.

Many business programmes also impose an admission-to-the-major step with its own grade point requirement, and a minimum grade in each core course. Confirm both against your own catalog year.

Course format and position in the curriculum

• Lecture with substantial project and applied work. ⚠ ISM courses are consistently more time-consuming than their credit value suggests, because software work expands to fill the debugging available.
• ⚠ No "permission is required" marking appears anywhere in the ISM prefix, and no fee notices — enrolment is gated by ISM3011 and by standing.
• Ask about delivery mode. UWF offers substantial online provision in the College of Business, and an online section of a project-based course demands more self-management, not less.
• UWF publishes no contact hours or terms of offering. A course offered in one term only will delay a sequence by a full year if missed — confirm with the department.

ISM4400 is 3 semester hours at the University of West Florida.