Document toolboxDocument toolbox

Working in the graph diagram

Overview

After the creation process, you will see the graph displaying all the connections between the distinct values in the columns you added. The graph is composed of nodes and links:

  • Node - Representation of each distinct value in the selected columns. Nodes may have different colors and labels (as set in the creation process) to clearly identify the values corresponding to each column. By default, nodes appear with a black label indicating their size. Hover over a node to see the type you assigned to it in the creation process, its size and its value. 
    By default, the size of a node represents the sum of all the links it has with other nodes, but sizes can be calculated using different criteria that you can set in the Graph diagram menu (more info in the Graph construction section below).

  • Link - Links represent the connections between values of the different columns added. The direction of the arrows is set during the creation process. 
    By default, the width of a link represents the number of occurrences of the pair of values it connects. However, if you added a metric in the graph creation process, you can set different criteria to set the width of each link (more info in the Graph construction section below).

Graph diagram menu

The panel to the left of the graph diagram contains a menu of options that you can use to customize the appearance of the graph, set specific metrics to calculate the size of your nodes and links, and display additional information to complement your graph.

The Graph diagram menu contains the following groups of settings:

Search specific nodes by their names. The search is case-insensitive and you may type the whole name or part of it. Hit the Enter key to highlight all the matching nodes in the diagram.

Apply filters to determine the nodes you want to display in the graph. Use the controls in each bar to set the minimum and maximum values of each filter type. You can filter by:

Link width

Filter nodes that have links with a specific width range. The minimum and maximum width values will vary depending on what you set in the Link widths field of the Graph construction section (see Graph construction section below).

Node size

Filter nodes within a specific size range. The minimum and maximum sizes will vary depending on what you set in the Node sizes field of the Graph construction section (see Graph construction section below).

Node degree

Number of connections that a node has with other nodes.

A note on filters

Note that filters should be applied in the order they are displayed on the menu: first by link width, secondly by size and finally by degree. Each time you apply a filter, all the weights are recalculated and new values are assigned to link widths and node sizes.

Suppose you filter by node size > 10, but in the final result you can see nodes with size 3. This is because after filtering by node size > 10, all the weights are recalculated and it is possible to have nodes that initially had a size > 10, but now have a size 3.

Apart from using the filter bars, you can select a node or a group of nodes and use the tools in the Selection area to hide them, or hide the unselected ones. You can also select links to filter the nodes they connect. To select multiple nodes or arrows, click on the white area and drag the mouse to define an area that contains the desired nodes or arrows.

To do this, the icon below the navigation arrows must be this . Click it again to turn it to this , which allows you to click on the white area and drag the whole graph.

Once you have selected the required nodes or arrows, you can apply the following filters in the Selection area.

Filter in the selected nodes and hide the rest of them.

Filter out the selected nodes and display the rest of them.

Filter out the selected links.

Undo a filter.

Redo a filter.

Clear all the filters applied.

Size and width automatically recalculated

Note that the size of the nodes and the width of the links will be recalculated after applying a filter.

If you switch on the Map mode toggle, all nodes with geo-coordinates are plotted on a world map. This is useful for analyzing the geographical distribution of the nodes.

After enabling this option, two other settings appear:

Title set

Choose a style for the map between Default, Terrain, Roads, Imagery, Light, Dark and Night.

Show unpositioned nodes

You can hide nodes with no geo-coordinates disabling the Show unpositioned nodes option. If the switch is toggled on, nodes that do not have geo-coordinates are grouped in the Atlantic Ocean by default. If needed, you can set a different area to display them entering a specific latitude and longitude values.

There are several options available that you can use to customize the way your graph is built:

Ignore nulls

If activated, nodes representing null values will not appear in the graph.

Convert null synonyms

If activated, it converts null synonyms (unknown, none) into null.

Link widths

Select the metric you want to use to calculate the link widths. By default, the option Count of link occurrences is selected, which considers the number of times the pair of values linked occur.

If you add a metric during the graph creation process, you will also be able to choose one of the following options:

  • Sum of metric_name field - The width of the links is calculated using the sum of the metric values corresponding to each pair of nodes.

  • Average of metric_name field - The width of the links is calculated using the average of the metric values corresponding to each pair of nodes.

  • Max of metric_name field - The width of the links is calculated using the maximum metric value for each pair of nodes.

  • Min of metric_name field - The width of the links is calculated using the minimum metric value for each pair of nodes.

Select the way node sizes are calculated using the options in the Node sizes drop-down menu. The size of a node is indicated in the black label that appears at the top of each node.

All links

Sum of all links

The sum of all the link widths, both incoming and outgoing.

Average of all links

The average of all the link widths, both incoming and outgoing.

Max of all links

The maximum width of all the node links.

Min of all links

The minimum width of all the node links.

Incoming links

Sum of incoming links

The sum of all the incoming link widths.

Average of incoming links

The average of all the incoming link widths.

Max of incoming links

The maximum width of all the incoming links.

Min of incoming links

The minimum width of all the incoming links.

Outgoing links

Sum of outgoing links

The sum of all the outgoing link widths.

Average of outgoing links

The average of all the outgoing link widths.

Max of outgoing links

The maximum width of all the outgoing links.

Min of outgoing links

The minimum width of all the outgoing links.

Graph properties

Betweenness

Nodes with a high betweenness level are the ones that most frequently act as ‘bridges’ between other nodes.

Closeness

This is the measure that helps you find the nodes that are closest to the other nodes in a network, based on their ability to reach them.

To calculate this, the algorithm will find each node’s shortest path to every other node, then assigns each node a size based on the sum of all the paths.

Eigencentrality

Eigencentrality is a measure of influence that takes into account the number of links each node has and the number of links their connections have, and so on throughout the network.

Page rank

The PageRank algorithm identifies important nodes by assigning each a score based upon its number of incoming links. These links are weighted depending on the relative score of its originating node.

[ 1 Selection mode ] [ 2 Timebar ] [ 3 Layout ] [ 4 Nodes ] [ 5 Links ] [ 6 Screen scale ] [ 7 Number format ]

Selection mode

This determines which related values will be highlighted when you double-click a node multiple times successively. The originally-selected node is considered as level 0. Double-click the level-0 node once to select the nodes that have a level 1 relationship with it, double-click again to also select the level-2 nodes, and so on.

Hops

When you double-click a level-0 node multiple times successively, the nodes that are within n hops from it will be selected.

See below how we double-click a node twice and all the nodes that are 2 hopes from it are selected.

Same row

This will select the nodes in the same row as the level 0 node. Double-click the node selected to keep on showing levels. 

Thread

This will select the nodes from the same type as the level 0 node. Double-click the node selected to keep on showing levels.

Tree

This will select the nodes the parent and child nodes of the level 0 node. Double-click the node selected to keep on showing levels.

When you select a node, links going from that node to other ones are green. Links arriving to that node are purple and discontinuous. 

Timebar

This powerful tool helps you interpret how your connected data evolves and behaves over time.

The histogram shows the overall amount of data, whilst the line shows the amount of the specific nodes you select in the graph. Select different groups of nodes using the CTRL key to display different lines and analyze them simultaneously.

Use the time bar sliders or the zoom controls to zoom to periods of interest and change the time range visualized in the graph. The navigation controls allow you to play, pause, fast-forward or rewind the time bar.

Click to set the default time bar zoom.

Zoom in the time bar. You can also double-click the time bar or use the mouse wheel.

Zoom out the time bar. You can also use the mouse wheel.

Rewind the time bar.

Click to play/pause the evolution of your data over time.

This alternative play button makes the left slider stay fixed, only the right one moves through time. This allows you to check how data "accumulates" through time. This is useful for checking, for example, how a virus expands geographically.

Fast forward the time bar.

See below how the graph is updated when you zoom in the time bar, and how nodes evolve when pressing the play button.

Layout

You can configure the graph arrangement by selecting the required type of layout, and defining its Orientation and Tightness.

 

 

Type

Standard

The graph links are a consistent length, nodes and edges overlap as little as possible and are distributed evenly in the graph.

Structural

This functions in a similar manner to the standard layout, but nodes with similar attributes are grouped together in fans.

Sequential

This layout is a good way to display data that contains a clear sequence of links between distinct levels of nodes. It minimizes link crossings and makes really efficient use of the available screen space.

Hierarchical

Child nodes are shown in horizontal layers below their parents, with the connections working top-down through the chart from the original subject.

Radial

Arranges nodes in concentric circles around the original subject in a radial tree.

Lenticular

Push highly-connected nodes into the center, and force the less connected nodes into the periphery.

Orientation

Choose between Default, Top to bottom, Bottom to top, Left to right or Right to left.

Tightness

Select the tightness level of the graph. Nodes to show the clustered closer or spread further apart. Values allowed are 1-10.

Nodes

Set options related to the graph nodes:

Show info

Activate this option to display the size of the nodes on the graph.

Hide singletons

Activate this option to hide the nodes with no links.

Truncate labels

Activate this option to truncate the length of node labels when they are excessively long.

Links

Configure the aspect of the graph links:

Show info

Activate this option to display the width of the links on the graph.

Curvature

Define the curvature of the links. The default value is 90 degrees.

Transparency

This defines the degree of transparency of normal and selected links. Enter a value from 0 (minimum) to 1 (maximum).

Screen scale

Set the default size of nodes and links in your graph.

Number format

Select the format of the node sizes and link widths, and set the required number of decimal places in the decimals field.

No format

Display numbers with no thousands separators.

Thousands separators

Display numbers with thousands separators.

SI suffixes (factor 1000)

Format values in thousands (K), millions (M)...

IEC suffixes (factor 1024)

Apply the IEC numbering system.

Scientific notation

Express numbers in scientific notation.

Duration

Express numbers as durations.