The Graph is the foundation for exploring data in Graphext. It contains a visual representation of your data points often grouped together into clusters. The position of your points - or nodes - on the Graph represents their relationship to one another. The nature and strength of their connections is defined by the recipe you create.
Use the Graph to begin investigating patterns in your dataset. You can customise its appearance, zoom in and examine specific groups or take a snapshot of it's layout and save it as an insight.
"Mankind invented a system to cope with the fact that we are so intrinsically lousy at manipulating numbers. It's called the graph."
- Charlie Munger
A node is a junction on the Graph representing one of your data points. Every row in your dataset is represented by a node and plotted on the Graph. The position of a node is defined by its features and their connection to the features of other nodes. These relationships are represented as links between nodes.
Use quantitative variables to add a size dimension to your Graph. Data points with a higher value for a set variable will be represented as larger nodes in your Graph. You can set a default value for your node sizes in your project's settings.
The larger a node is, the higher the value it has for a set variable.
If you haven't activated node size mapping, you can change the size used to represent your nodes using a one-dimensional scale. Click the node size icon from the top of your Graph and move the slider up and down to adjust the sizes of nodes.
You can also complete this action using the Graph tab inside your project settings window. Use this course of action to save your preferred node size.
When you apply node size mapping to your data, you can control the range of sizes presented in your Graph. Click the node size icon from the top of your Graph and drag the upper and lower points to change the range of node sizes presented.
Regions are areas of your Graph defined by the presence of a number of values within a variable. Regions can be defined by any variable in your data. When you set region labels, the names of values featured within that variable will appear in your Graph over the areas featuring groups of data points with that value.
Change the variable representing your region labels using the project settings window.
Labels are the text that appears over nodes. They show a node's value for a particular variable. You can change what variable a label represents in your project settings.
When you search for data in the Graph, your query is matched against node labels.
Adjusting the density of labels will change how many node labels appear. Zooming into the Graph makes more labels appear.
They show a node's value for a particular variable. Adjusting the density of labels will change how many node labels appear.
Links between nodes represent the connection between rows in your dataset. Links define a node's position by pulling it towards nodes with similar features and therefore creating clusters. Hiding node links can sometimes create a cleaner aesthetic inside your Graph and allow for easier recognition of smaller clusters.
The position of nodes in the Graph was defined in the setup of your project. Hiding the links between nodes won't change their position.
Color is an immediately recognisable feature of design distinguishing one group from another. Colour mapping is a powerful way to signal the difference between different groups in your data. Using the variable charts in the left and right sidebars you can apply colour mapping to any feature of your dataset.
Emphasize data groups by applying colour mapping to new segmentations that you create. This also changes the color of your node links.
Alongside using the representations of your variables in the sidebar to filter your data, you can directly select data in the Graph, select a point and its neighbors or search for specific data.
Use the visual representation of your data to find patterns then zoom in to examine them in greater detail. Selecting a group of data points from the Graph allows you to conduct your analysis in greater detail.
Selecting nodes directly in the Graph means that you can react quickly when you spot new patterns. Use the direct selection tool to click and drag your mouse over nodes. You can save these selections for later or keep them as insights.
Click outside of the direct selection to remove it and return to the whole dataset.
Find important nodes in your Graph then identify linked data points sharing their significant features. For instance, selecting the neighbors of a node can help you find customers sharing characteristics with a key customer you are analysing.
A node has neighbours because Graphext calculates links between data points based on their similarities.
Searching for data in the Graph means that you can quickly find specific nodes or groups that share similar features. The search tool matches your query with the tags that you gave your data when setting up your project.
Tags are the text that appears over groups of nodes. You can change what variable a tag represents in your project settings.
You can perform a number of actions on your Graph or Graph selections. These include deleting selections, clustering selections, saving selections as well as downloading a screenshot or capturing your Graph or as an insight.
To learn more about Saving, Deleting and Clustering Selections, head here.
You can download a screenshot of your full Graph or after having made a selection. Screenshots will download as 'png' image files, which you can then edit yourself, send on to your teammates or include in a report.
The screenshot tool will take a snapshot of how the Graph currently appears in your project.
Capturing an Insight from the Graph will save the current visual representation of your data to an insight card. You can use insights to share your analysis with others by presenting directly inside of the Insights panel or by exporting insights either individually or collectively.
Pressing the play icon at the bottom of an insight allows you to jump straight back into the point of your analysis at which you saved that insight.
We know that data isn't always clean and simple.
Have a look through these topics if you can't see what you are looking for.