Glossary

Key terms and concepts in Flow Myna and process mining.

Process Mining Terms

Process Mining

Data-driven technique for discovering, monitoring, and improving business processes by analyzing event logs.

Event

A single occurrence of an activity at a specific point in time. Example: "Block Sales Order on SO-000031 at 2025-02-13 09:56."

Activity

A type of work performed in the process. Also called Event Type. Example: "Create Sales Order," "Block Sales Order," "Clear Invoice."

Case

A single instance flowing through the process. In a case-based view of order to cash, one sales order, such as SO-000031.

Object

An entity in your process. In object-centric mining, multiple object types can exist (Sales Orders, Sales Order Items, Deliveries, Invoices, Customers). Traditional mining uses one object type called "case."

Event Log

Historical record of events that occurred, typically exported from business systems. The source data for process mining.

Process Map

Visual node-and-edge diagram showing all activities and transitions in your process with performance metrics.

Variant

A unique sequence of activities that cases follow through the process. Different execution paths.

Conformance

How well actual process execution matches the intended or designed process.

Bottleneck

An activity or transition that significantly slows down the process. Identified by long duration times.

Rework

When cases loop back to repeat previous activities, typically due to errors or rejections requiring corrections.

Flow Myna Concepts

Workspace

Your central hub for all process data and team collaboration. Like a data warehouse containing multiple datasets and projects.

Dataset

Uploaded process data containing objects, events, and relationships. Created by the upload wizard.

Project

A focused analysis workspace that selects specific object types and event types from your datasets.

Data Scope

The selection of which object types and event types are included in a project's analysis.

Agents

The pipeline that reads an upload and works out what it means: Reader parses the files, Mapper works out the object and event model, Inflator expands it for review. You approve the result before anything is imported.

Co-Pilot

Conversational assistant that answers questions about the process in a project, applying filters, running analyses and drawing charts from your data.

Filter

A condition that narrows the analysis to a cohort. Several filters combine with AND, and every page in the project answers to the same set.

Saved filter

A filter given a name and description, stored with the project so anyone with access can apply it.

Pinned filter

One saved filter promoted to the project's baseline: always applied, on every query, for everyone. At most one per project, and only a project admin can set it.

Crop

A transformation that trims every object's timeline to the segment between two activities, so the map and its numbers cover only that stretch. Unlike a filter, it keeps the objects and shortens their timelines.

Terminal event

An event type that marks the process finished. Terminal events are what make "completed", "in-flight" and cycle time meaningful, and are set in Data Scope.

Enrichment

A derived attribute computed into the data once from a formula — days late, an invoice's due date — then usable in every filter, chart and answer. See Enrichments.

Metric definition

A named business metric with an agreed formula, confirmed in conversation and promoted by an admin, so the same question returns the same number for everyone. See Metrics.

Rule

A statement of what should never happen in the process, with a count of how often it does. An SLA target is a rule with a threshold.

Board

A project's dashboard of pinned charts, re-run against current data on demand and answering to the project's active filters.

Object-Centric Mining Terms

Object Type

A category of entities in your process. Example: "Sales Order," "Sales Order Item," "Invoice."

Object-Event Relationship

Connection between an object and an event. Many-to-many: one event can relate to multiple objects, and one object can experience many events.

Object-Object Relationship

Connection between different object types. Example: a Sales Order belongs to a Customer, and each Sales Order Item belongs to a Sales Order.

Default Lens (Primary Object Type)

The object type a project's process map is centred on by default. When a project includes several object types, you can switch the lens to view the same process from a different object's perspective.

Interleaved Events

Events from a related object type overlaid on the current lens's process map to give context — for example, showing an order's delivery and invoice events alongside its own flow.

Parallel track

A lane of activities that runs concurrently with other lanes between the same two milestones. Tracks can be detected from the data or declared by hand; the map draws each in its own colour with dashed fork and rejoin edges.

Critical path

The parallel track the process is actually waiting on at the rejoin. Shortening any other track buys nothing.

Variant signature

The ordered sequence of activities that defines a variant, used as a durable filter: "the objects that follow exactly these paths", which stays meaningful as new data arrives.

Statistical Terms

Mean (Average)

Simple average of all values. Sum divided by count.

Median

Middle value when sorted. 50% of cases are faster, 50% slower. Less affected by outliers than mean.

P95 (95th Percentile)

Value at which 95% of cases are faster. Used to understand slower cases and set SLA targets.

Standard Deviation

Measure of variability. High = inconsistent timing, Low = predictable process.

Duration

Time between two events or total time from start to end of process.

Frequency

How often something occurs. Usually expressed as count or percentage.

Throughput

Number of cases completed per time period (cases/day, cases/month).

Cycle Time

How long one object took: from its first activity to its completion if it finished, or to the latest activity in the data if it is still open. This is what Flow Myna's duration tiles, charts and duration filter measure.

Service time

For an event that records both a start and an end, how long the work itself took. Shown on nodes.

Waiting time

The gap between one activity and the next. Shown on edges. Most process questions turn on the difference between this and service time.

Arrivals

The count on the left of a node: how many times the activity happened for the objects in view, repeats included. It equals the sum of the counts on the arrows entering the node. When some objects arrived more than once, the node panel adds Objects, the distinct count.

Remaining

The count on the right of a node: objects that reached this activity and have not moved on. Where work is sitting now, as opposed to what happened historically.

Process Patterns

Happy Path

The intended, standard execution sequence. Usually the most common variant.

Anomaly

Unusual pattern that deviates significantly from normal execution. May indicate errors or special cases.

Exception Handling

Alternative paths taken when errors occur or special circumstances arise.

Parallel Processing

When several activities happen at the same time or in no fixed order. Modelled explicitly as parallel tracks, so that re-orderings within a lane are not counted as different variants.

Sequential Processing

When activities must occur in strict order, one after another.

Performance Terms

SLA (Service Level Agreement)

Target performance threshold. Example: "Ship 90% of orders within 7 days."

Utilization

How busy a resource is. High utilization may indicate capacity constraints.

Queue Time

Time spent waiting before an activity begins. Often indicates bottlenecks.

Processing Time

Active time spent performing an activity (excluding wait time).

Lead Time

Total time from request to delivery, including all waiting and processing.

Data Terms

CSV (Comma-Separated Values)

Text file format for structured data. The most common upload format, and the one the mapping agents are built for.

XES

IEEE 1849, a standard event-log interchange format. Uploaded directly; skips the mapping step because the structure is already declared.

OCEL 2.0

Object-Centric Event Log, the interchange standard for logs with several object types. Flow Myna both imports OCEL 2.0 (.jsonocel, SQLite) and exports any dataset as OCEL 2.0, so your data is never locked in.

Timestamp

Date and time when an event occurred. Required for all events.

Attribute

Additional information about an object or event. Examples: amount, category, risk level, region.

Case ID

Unique identifier for a case in traditional process mining. In object-centric mining, use object IDs instead.

Foreign Key

Column that references another entity. Used to establish relationships.

Analysis Terms

Root Cause Analysis

Investigation to identify underlying reasons for problems or patterns.

Cohort

Group of similar cases. Example: "Blocked orders" or "Orders from the UK."

Comparison

Analysis technique showing differences between two or more cohorts.

Drill-Down

Analyzing at increasing levels of detail. Start broad, then narrow focus.

Roll-Up

Aggregating detailed data to higher-level summary. Opposite of drill-down.

Segmentation

Dividing cases into groups based on attributes for separate analysis.

Quick Reference

Want to analyze your process?

1. Create workspace
2. Upload data (dataset)
3. Create project
4. View process map
5. Analyze variants
6. Ask the Co-Pilot

Need help with a term?

• Use search (Cmd/Ctrl + K)
• Ask the Co-Pilot
• Check relevant documentation page
• Contact support