Process mining and analytics engagement for an enterprise client's internal procure-to-pay system. This multi-week engagement focused on evaluating process compliance and identifying optimization opportunities through data-driven analysis.
Note on Data: All identifying details, exact metrics, and proprietary references have been generalized to protect client confidentiality. Comparative findings, severity levels, and order-of-magnitude impacts are preserved to reflect analytical validity.
- Project Objectives
- Technical Approach
- Key Findings
- Business Impact & Recommendations
- Technical Skills Demonstrated
- Documentation
- Contact
- Achieve end-to-end procurement process visibility
- Identify bottlenecks, inefficiencies, and compliance risks
- Validate effectiveness of simplified workflows for low-value purchases
- Uncover data quality issues affecting process management
| Category | Tools |
|---|---|
| Data Processing | Python (pandas, numpy) |
| Process Mining | ProDiscovery |
| Data Visualization | Matplotlib, Process Flow Diagrams |
| Data Sources | Procurement System, Workflow/Approval System, Financial Settlement System |
Objective: Transform disparate system data into unified event log
Activities:
- Standardized field names and timestamp formats across multiple source systems
- Implemented attribute-based mapping to connect procurement, approval, and payment records
- Generated missing timestamps to enable continuous process flow analysis
- Applied data quality filters and validation rules
Output:
- Filtered event log: hundreds of cases, 1,200+ events, 15+ distinct activities
- Event coverage: ~70% of target process steps
- Majority of cases met analysis criteria after quality filters; remainder excluded for completeness
Objective: Reconstruct and analyze actual process flows
Activities:
- Event log reconstruction from multi-system data
- Process conformance checking against standard workflow
- Process pattern identification and deviation analysis
- Lead time decomposition (processing time vs. waiting time)
- Dotted chart analysis for temporal pattern recognition
Output:
- Standard process model vs. actual process variants
- Conformance rate calculation
- Deviation pattern taxonomy (3 types identified)
Objective: Generate actionable insights and recommendations
Activities:
- Statistical correlation analysis: transaction size vs. processing speed
- Root cause analysis for long-duration outlier cases
- Organizational-level approval delay investigation
- Simplified workflow effectiveness validation
Output:
- Multiple analytical scenarios completed
- Quantified business impact metrics
- Strategic roadmap for process mining expansion
Summary: Majority of procurement cases follow standard process, with identifiable deviation patterns
Detailed Findings:
- ~80–85% of cases followed the standard procure-to-pay workflow (request → selection → order → receipt → inspection → payment).
- ~15–20% showed process deviations categorized into three distinct patterns:
- Type A: Orders initiated before request registration (sequence violation)
- Type B: Requests registered after goods receipt (retrospective logging)
- Type C: Contract finalization after order issuance (out-of-sequence approval)
Implications:
- High compliance risk in deviation cases
- Need for pre-submission validation controls
- Audit exposure in non-conforming transactions
Summary: Simplified workflows demonstrate significant cycle time reduction
Detailed Findings:
- Simplified approval process achieved ~80–90% cycle time reduction vs. complex path:
- Streamlined path: median cycle time in the order of hours (single-digit to low double-digit)
- Complex path: median cycle time in the order of days to weeks
- Strong inverse correlation validated: lower transaction amounts → faster processing
- Approval step count directly impacts total lead time:
- Single-step approval: fastest processing
- Multi-step approval: proportional delay increase
Implications:
- Low-value purchase simplification initiative is highly effective
- Opportunity to expand simplified workflows to more categories
- Potential for automated approval thresholds
Summary: Significant portion of cases experience extended lead times driven by approval delays
Detailed Findings:
- ~35–45% of cases exceeded average lead time
- Lead time distribution: mean in the order of months; median slightly lower but still order of weeks to months
- Root cause identified: Approval waiting time (primary driver), especially at payment/financial sign-off stage
- Secondary factors:
- Organizational-level approval variations
- Multi-department coordination requirements
- Certain organizational units showed consistently longer approval times
Implications:
- Approval stage requires process redesign (high priority)
- Organization-specific training or staffing adjustments needed
- Opportunity for threshold-based automation
Summary: Critical timestamp data missing in source systems, requiring complex workarounds
Detailed Findings:
- Critical gap: Native event timestamps missing in core transactional and financial systems
- Impact: Process mining not feasible without cross-system integration and derived timestamps
- Workaround implemented: Attribute-based mapping using workflow/approval system as bridge
- Achieved ~70%+ event coverage
- Required non-trivial data transformation logic
- Manual validation and reconciliation needed
Implications:
- Current architecture not optimized for process analytics (high-risk for scalability)
- Enterprise-wide event logging standards required
- Significant manual effort needed for ongoing analysis without remediation
Action: Expand simplified approval workflows to additional low-value purchase categories
Implementation:
- Increase transaction amount threshold for simplified process
- Reduce approval steps for routine, low-risk purchases
- Implement categorical exemptions (e.g., routine supplies, standard services)
Expected Impact:
- ~30–40% reduction in overall procurement lead time
- Reduced administrative burden on approval stakeholders
- Improved user satisfaction with procurement process
Timeline: 0–3 months
Action: Implement automated approval thresholds for routine purchases
Implementation:
- Define rule-based auto-approval criteria (amount, category, vendor)
- Deploy pre-approved vendor framework for recurring purchases
- Establish escalation protocols for exceptions
Expected Impact:
- Address ~35–45% of long-duration cases
- ~50%+ reduction in approval waiting time for eligible transactions
- Free up approval capacity for high-value/high-risk transactions
Timeline: 3–6 months
Action: Deploy pre-submission validation checks to reduce non-conformance
Implementation:
- System-enforced workflow sequence validation
- Mandatory field completion before advancement
- Real-time compliance alerts to users
Expected Impact:
- Reduce non-conformance rate from ~15–20% to <5%
- Minimize audit exposure and compliance risk
- Improve data quality and completeness
Timeline: 3–6 months
Scope: End-to-end customer-facing process (e.g., inquiry-to-conversion journey)
Key Metrics:
- Conversion rate and lead time by segment
- Drop-off points in funnel
- Customer satisfaction correlation with process efficiency
Expected Value:
- Identify friction points in conversion funnel
- Optimize customer touchpoints
- Increase conversion rate and revenue
Prerequisites:
- Customer interaction event logging
- Integration with CRM and transactional systems
Scope: Fraud detection, audit compliance, information security
Key Metrics:
- Abnormal transaction detection rate
- Policy adherence by process and department
- Risk incident occurrence and resolution time
- Audit finding trends
Expected Value:
- Real-time risk identification and mitigation
- Proactive compliance management
- Reduced audit findings and regulatory exposure
Prerequisites:
- Risk event taxonomy definition
- Integration with security and audit systems
Scope: Recruitment, onboarding, performance management
Key Metrics:
- Time-to-hire by role and department
- Onboarding completion rate and duration
- Employee turnover drivers and patterns
- Promotion and advancement cycle times
Expected Value:
- Improve talent acquisition efficiency
- Enhance employee retention
- Data-driven HR policy optimization
Prerequisites:
- HR system event logging capability
- Privacy and compliance framework for employee data
Enterprise systems lacked consistent event logging, preventing native process mining capability and requiring significant manual intervention for analysis.
1. Event Logging Standards
- Define enterprise-wide event schema (case ID, activity, timestamp, actor, system)
- Implement standardized logging across business-critical systems
- Establish event granularity requirements by process type
2. Data Quality Framework
- Completeness: All process steps must generate events with required attributes
- Timeliness: Events logged in real-time or near-real-time (e.g., <1 hour delay)
- Consistency: Standardized field formats, naming conventions, and data types
- Accuracy: Validation rules to prevent invalid or contradictory data entry
3. Automated Data Validation
- Real-time data quality monitoring dashboards
- Automated alerts for missing or anomalous events
- Scheduled data reconciliation jobs across systems
4. Cross-System Integration Framework
- Master data management for common entities (vendors, employees, departments)
- API-based event streaming between systems
- Centralized analytics layer for process mining
| Phase | Timeline | Deliverables |
|---|---|---|
| Assessment | Month 1–2 | Current state analysis, gap identification, stakeholder alignment |
| Design | Month 3–4 | Event logging standards, data model, integration architecture |
| Pilot | Month 5–7 | Implement in 1–2 pilot systems, validate approach |
| Rollout | Month 8–12 | Broader implementation, training, change management |
| Optimization | Ongoing | Continuous improvement, expanded analytics use cases |
- Process Visibility: Real-time monitoring of business processes
- Continuous Improvement: Data-driven optimization culture
- Scalability: Foundation for AI/ML-based process automation
- Compliance: Audit trail and regulatory compliance capability
| Skill | Application in Project |
|---|---|
| Event Log Reconstruction | Built unified event log from multiple disparate systems with ~70%+ coverage |
| Process Conformance Checking | Validated ~80–85% adherence to standard workflow, identified multiple deviation types |
| Root Cause Analysis | Identified approval stage as primary driver of ~35–45% of long-duration cases |
| Statistical Correlation Analysis | Proved inverse relationship between transaction size and processing speed |
| Process Pattern Recognition | Classified non-conforming cases into distinct violation taxonomies |
| Temporal Analysis | Decomposed lead time into processing vs. waiting time components |
| Skill | Application in Project |
|---|---|
| Multi-System Integration | Integrated procurement, workflow, and financial systems with different schemas |
| Data Mapping & Transformation | Implemented attribute-based matching for cross-system record linkage |
| Timestamp Reconstruction | Generated missing event timestamps using workflow/approval system timestamps |
| Data Quality Assessment | Applied completeness criteria; filtered to analyzable subset from larger source population |
| Event Schema Design | Standardized field names, formats, and data types across heterogeneous sources |
| Data Validation Logic | Removed duplicates, invalid activities, and non-identifiable records |
| Skill | Application in Project |
|---|---|
| Requirement Gathering | Defined analytical scenarios aligned with stakeholder priorities |
| KPI Framework Design | Proposed industry-aligned KPIs for enterprise value chain |
| Strategic Roadmap Planning | Developed multi-phase expansion plan with timeline and resource requirements |
| Stakeholder Communication | Presented findings with fact-checking validation sessions |
| Business Case Development | Quantified ROI for workflow simplification and automation initiatives |
| Change Management | Recommended data governance framework with implementation roadmap |
| Category | Tools | Proficiency Demonstrated |
|---|---|---|
| Process Mining | Process mining software | Discovery, conformance checking, variant analysis, temporal charts |
| Data Processing | Python (pandas, numpy) | Data cleaning, transformation, merging, statistical analysis |
| Data Visualization | Matplotlib, process diagrams | Cycle time charts, correlation plots, process flow maps |
| Systems | ERP / procurement / workflow / financial platforms | Cross-system data extraction, schema mapping |
- Sample case documentation (see repository for available assets) — Analysis overview with approach and generalized findings.
process-mining-portfolio/
├── README.md
├── .gitignore
└── docs/
└── Process_Analytics_Portfolio_Sanitized.pdf
This portfolio illustrates work from a process analytics engagement with an enterprise client. Identifying details, exact metrics, and proprietary references have been generalized; comparative insights and impact magnitudes are preserved for portfolio purposes.
Minwoo Park
- Email: Minwoo.Park219@gmail.com
- Phone: +1 (706) 461-3489
- LinkedIn: linkedin.com/in/mp74484
Last Updated: January 2026