PM4PY Suite

PM4PY Suite is a modern desktop UI for Process Mining with PM4PY, built with Flet.
It provides a comprehensive graphical interface for importing event logs, discovering process models, performing conformance checking, analyzing variants, and exporting results — all without writing code.

The application wraps many core PM4PY features into an easy-to-use UI designed for analysts, researchers, and students.

---

Features

Event Log Handling

• Import XES, CSV, and Parquet event logs
• CSV column mapping for case, activity, and timestamp
• Log statistics and overview

Process Discovery

• Alpha Miner (Classic / Plus)
• Inductive Miner (IM / IMf / IMd)
• Heuristics Miner
• Directly-Follows Graph (DFG) discovery

Conformance Checking

• Token-based Replay
• Alignments-based Conformance Checking
• Load external models (PNML, BPMN) or use discovered models

Log Analysis

• Variant analysis (Top process variants)
• Event log statistics
• Performance analysis (waiting times)
• Social network analysis of resources

Filtering

• Time range filtering
• Case size filtering
• Top-K variant filtering

Simulation

• Simulate event logs from discovered Petri nets

Export

• Export event logs:
  • XES
  • CSV
  • Parquet
• Export process models:
  • PNML
  • BPMN

---

Technologies Used

• Python
• Flet – Modern Python UI framework
• PM4PY – Process mining library
• Pandas – Data processing
• PyArrow – Parquet support

---

Installation

1. Clone Repository

git clone https://github.com/yourusername/pm4py-suite.git
cd pm4py-suite

2. Install Dependencies

pip install flet pm4py pandas pyarrow

---

Run the Application

python app.py

Flet will start the UI and open the application in your browser or desktop window.

---

Supported Event Log Formats

Format	Description
XES	Standard process mining event log format
CSV	Custom event logs with column mapping
Parquet	Efficient columnar data format

Required Columns

Column	Meaning
case:concept:name	Case identifier
concept:name	Activity name
time:timestamp	Event timestamp

Optional:

Column	Purpose
org:resource	Required for Social Network Analysis

---

UI Modules

Module	Description
Log Import	Load event logs
Alpha Miner	Classic discovery algorithm
Inductive Miner	Sound process discovery
Heuristics Miner	Frequency-based discovery
DFG Discovery	Directly-Follows Graph
Token Replay	Basic conformance checking
Alignments	Optimal conformance checking
Log Filtering	Filter traces and events
Variant Analysis	Top process variants
Statistics	Event log metrics
Social Network	Resource interaction analysis
Performance	Waiting time analysis
Simulation	Generate synthetic event logs
Export	Save logs and models

---

Example Workflow

1. Import an event log
2. Run Inductive Miner
3. Inspect the Petri net
4. Run Token Replay
5. Analyze Variants
6. Export the resulting BPMN or PNML model

---

Requirements

• Python 3.9+
• PM4PY compatible environment

Tested with:

• Python 3.10
• PM4PY latest version
• Flet 0.82+

---

Known Limitations

• Alignments may take several minutes on large logs
• Social network analysis requires org:resource
• Very large logs (>1M events) may require more memory

---

License

MIT License

---

Acknowledgements

• PM4PY Team https://pm4py.fit.fraunhofer.de/

• Flet Framework https://flet.dev/

---

Contributing

Pull requests and improvements are welcome.

Possible improvements:

• OCEL support
• Additional discovery algorithms
• Interactive process visualization
• Dashboard charts
• Docker deployment

---

Author

Process Mining UI built on top of PM4PY using Flet.