Skip to main content

Celonis Product Documentation

Network Explorer and Signal Link Explorer

  • The Network Explorer (NE) is a component for Views that allows the visualization and manipulation of data as a graph displaying the relationships between the data points as nodes connected to other nodes by edges, each with their respective KPIs and/or attributes.

  • The Signal Link Explorer (SLE) is a sister widget of NE that is intended to be used for the scenarios where "Signal Link" technology is available for the data model. The main difference to NE is the way it is configured and that it unlocks extra filtering options (see filtering).

Data structure
60362216.png

Output from tables defined on the right.

The network explorer requires two main separate tables on the data model to be able to work as expected, otherwise users might find unexpected behaviors (i.e some nodes not showing up, no data found after filtering, etc) when interacting with internal and external filters.

Nodes table: Mainly used as a lookup table, it helps to know which nodes the network contains, and their respective KPIs or attributes.

node_id: STRING

node_kpi: FLOAT

A

1

B

2

C

2

D

2

E

3

-A

1

Connections (i.e edges) table: Used to establish connections between nodes, and to assign KPIs or attributes to those relationships. Please note that both source and target must make reference to the node id defined in the table above, therefore we know that the node "A" has two relationships ("A" => "B" and "A"=>"C").

edge_source: STRING

edge_target: STRING

edge_KPI: FLOAT

A

B

1

A

C

1

B

D

2

C

E

3

-A

B

0

Configuration
60362229.png

NE using displayContentRules for nodes and edges.

YAML configuration

id: caseNetwork
type: ninja-explorer
settings:
  source: SOURCE
  target: TARGET
  nodes:  NODES
  edgeLabels:
    - field: EDGE_KPI
      displayContentRule: EDGE_COLORING  # add a display content rule to the first label to color the edge accordingly
  nodeLabels:
    - field: NODE_KPI
      displayContentRule: NODE_COLORING  # add a display content rule to the first label to color the node accordingly
  layout: HIERARCHY

displayContentRule (Defined in the knowledge model): Used to style nodes and edges with colors given a set of predefined rules.

...
customObjects:
  - id: NODE_COLORING
    displayName: Mode KPI display rule
    customAttributes:
      - comparator: GREATER_THAN
        value: 2
        style:
          backgroundColor: red
      - comparator: LESS_THAN_OR_EQUAL
        value: 2
        style:
          backgroundColor: green

Note

We currently only apply the displayContent rules defined in the first item of nodeLabels and edgeLabels. This is a current limitation of the component.

Filtering
60362243.png

NE node popup with basic filtering options

  • Shortest and longest Path: Is the basic filtering options provided for both NE and SLE. Given a graph the "shortest path filter" will find the path with least distance and the longest the opposite, the path with most distance. Distance in this context is determined by the weights (i.e KPIs) on both edges and nodes and NOT the actual geometrical longitude of the paths or the number of nodes and edges along it:

Shortest path

Longest path

60362233.png
60362232.png

Operation: 2.00 + 2.00 + 2.00 + 0.00 + 1.00 = 7.

Operation: 2.00 + 2.00 + 2.00 + 1.00 + 1.00 = 8.

Note: It is important to clarify, that in order for Shortest and longest Path filters to be calculated the the first item for nodeLabels and edgeLabels must be a KPI (i.e a FLOAT and not a STRING). This is a current limitation of the component.

Layouts
60362246.png

SNL using SPRING_FDP as the layouter

We offer different layout options to choose from.

  • CIRCLE: A circular layout.

  • HIERARCHY: Hierarchical layouting.

    • HIERARCHY_BT

    • HIERARCHY_LR

    • HIERARCHY_RL

  • SPRING_FDP:Based on spring force algorithms. Use for big networks.

  • SPRING_NEATO: Based on spring force algorithms. Use for small networks.