Practical Application

Apply your knowledge with hands-on exercises

Business Requirements Analysis Workshop

1
2
3
4

4-Step Requirements Workshop

Master business requirements analysis

Before diving into dimensional modeling, you must understand the business requirements. This workshop guides you through stakeholder analysis and requirement gathering techniques.

1

Stakeholder Analysis

Identify Users

Identify who will use the dimensional model and what they need.

👨‍💼
Chief Financial Officer (CFO)

Executive Leadership

Key Questions
  • "What's our revenue per patient by department?"
  • "How do costs vary by patient type?"
  • "Which departments are most profitable?"
  • "What's our capacity utilization?"
Frequency:Monthly, Quarterly
👩‍⚕️
Operations Manager

Hospital Operations

Key Questions
  • "What's the average length of stay by department?"
  • "Which physicians have the highest patient loads?"
  • "How do we optimize room assignments?"
  • "What are our peak admission times?"
Frequency:Daily, Weekly
📊
Quality Assurance Director

Patient Care Quality

Key Questions
  • "What's our patient satisfaction by department?"
  • "How do readmission rates vary by physician?"
  • "Which treatments have the best outcomes?"
  • "Are we meeting quality benchmarks?"
Frequency:Weekly, Monthly

Decisions & KPIs

  • Write down 2-3 concrete decisions each persona makes (e.g., budget approvals, staffing levels).
  • Capture the KPIs they use to judge success (LOS, readmissions, revenue, utilization).

Pain Points & Gaps

  • Note conflicting numbers between systems or reports that erode trust.
  • List key questions they cannot answer today, where the current data falls short.

Latency & History

  • Ask how fresh the data must be (daily, hourly, near real-time) for each persona.
  • Clarify how many years of history are needed for trending and regulatory reporting.
2

Business Process Identification

Prioritize

Map the operational processes that generate analytical data.

ProcessFrequencyData VolumeBusiness ValuePriority
Patient Admission500/dayHighCriticalHigh
Surgery Scheduling150/dayMediumHighMedium
Medication Dispensing2000/dayHighMediumMedium
Equipment Maintenance50/weekLowMediumLow

Selected Process: Patient Admission

Highest combination of frequency, volume, and business value

Candidate Grains

  • One row per admission event (per patient, per encounter).
  • One row per patient-day in the hospital.
  • One row per bed assignment (per bed, per day).

Fact Types & Bus Hints

  • Admission events – likely a transaction fact shared across Finance and Operations.
  • Daily census by ward – candidate periodic snapshot fact for capacity reporting.
  • Admission-to-discharge journey – potential accumulating snapshot fact.
3

Requirements Gathering Techniques

3 Methods

Learn structured approaches to extract business requirements that will drive your dimensional model.

Stakeholder Interviews

Essential Questions

  • "What decisions do you make daily/weekly/monthly?"
  • "What data do you currently use to make these decisions?"
  • "What questions can't you answer with current reports?"
  • "How do you measure success in your role?"

Pro Tips

  • Focus on business outcomes, not technical solutions.
  • Ask for specific examples and scenarios.
  • Clarify the frequency and urgency of each need.
Modeling implications: highlight the decisions, measures, and breakdowns mentioned here; they point directly to candidate facts (admissions, revenue, LOS) and shared dimensions (Patient, Provider, Location).
Business Rules Workshop

Key Areas to Explore

  • How are patients classified or segmented?
  • What defines a "successful" admission?
  • How are costs allocated across departments?
  • What are the key performance indicators?

Expected Output: clear business rules that will guide dimensional model design decisions.

Modeling implications: define conformed dimensions (Patient, Provider, Department, Payer) and flags that will drive slowly changing attributes and consistent KPIs.

Current State Analysis

Analysis Focus

  • Review existing reports and dashboards.
  • Identify data gaps and pain points.
  • Document current data sources and quality issues.
  • Understand manual processes and workarounds.

Why this matters: understanding current limitations helps design a dimensional model that solves real problems.

Modeling implications: surface data quality issues, missing business keys, and source system constraints that must be addressed in the Kimball bus architecture.

4

Requirements Documentation

Final Output

Structure and prioritize the gathered requirements so they can directly feed the dimensional design.

Business Objectives
  • Improve patient flow and reduce wait times.
  • Optimize resource allocation across departments.
  • Enhance financial performance monitoring.
  • Support quality improvement initiatives.
Key Business Questions

CFO Questions

  • What's our revenue per patient by department?
  • How do treatment costs vary by patient type?
  • Which services are most profitable?

Operations Questions

  • What's the average length of stay by department?
  • How can we optimize physician schedules?
  • Which rooms have the highest utilization?
Required Analytics
  • Patient admission trends by time period.
  • Length of stay analysis by department and physician.
  • Resource utilization rates.
  • Financial performance by service line.
  • Quality metrics and patient satisfaction.
Final Requirements Summary

Primary Business Process: Patient Admission and Treatment

Key Stakeholders: CFO, Operations Manager, Quality Director

Critical Success Factors: Financial performance, operational efficiency, quality metrics

Chosen Grain (preliminary): one row per patient admission event (per patient, per encounter) to support patient- and stay-level analysis.

History & SCD Focus: retain historical Patient, Provider, Department, and Payer attributes so reports reflect values at the time of admission.

Non-functional Needs: daily refresh by 6am, at least 5 years of detailed history, and trusted, reconciled metrics across Finance and Operations.

Next: Apply the four-step Kimball methodology to design the dimensional model.