The September 2021 release of Kenelyze is now live!
Kenelyze allows users to visualize their connected data on the fly in the form of fully interactive network maps. From this release onwards, it is now also possible to generate customizable matrix visualizations for any network created in Kenelyze. Additionally, the platform can now also ingest datasets formatted as matrices and visualize them as networks, as well as export any matrix visualization created in Kenelyze to allow for further external analysis.
Matrix visualizations
Matrix visualizations are based on the same source data as networks: nodes and the links between them. In matrices, nodes are placed along rows and columns, with links between nodes shown as cells at the intersection of rows and columns. When allowing a user control of how matrices are sorted and colored, they can be a powerful tool to uncover additional insights from connected data. They are especially useful for data which is densely connected. Network visualizations of this type of data are often hard to interpret because of the many links on screen. Matrices do not have this limitation, since links between nodes are visualized as cells at intersections. The key benefit of this is that matrices allow to uncover structure in your dataset, even when you’re working with highly connected data.
New features
Here is an overview of the new features:
Matrix View
Any small to medium-sized network can be visualized as a matrix using the new Matrix View button in the menu bar:
Matrix visualization is recommended for networks up to 500 nodes. If your network consists more than 500 nodes, Kenelyze will still be able to create a matrix, but will notify you that your network may be too large to create a usable matrix visual.
As soon as it is created, you will see a matrix which is sorted based on node labels and colored based on link weights. Nodes can be searched using the search bar at the top left. Clicking a search result then highlights the location of the node in the rows and columns. Hovering matrix cells shows more information at the bottom left; double-clicking a cell brings you back to the underlying network and highlights the location of the link and its connected nodes.
You can adapt the matrix to your liking by using the controls in the panel on the right. Columns and rows can be sorted based on any attribute associated with them. When sorting based on a textual attribute, Kenelyze will sort from largest group of nodes to smallest group of nodes associated with that attribute (as in the example above). Within groups, nodes are sorted based on their Degree (number of connections). When sorting based on a numerical attribute, nodes in rows and columns will be ordered from high (left/top) to low (right/bottom).
Label colors can be based on any attribute associated with nodes; in the same vein, cell colors can be colored based on any link attribute. You can adapt the size of labels and cells by using the sliders.
The controls just below the matrix allow you to zoom in, zoom out, recenter the camera and take a snapshot. The download button converts the matrix to a CSV file and keeps intact any row/column sorting you have applied.
Importing Matrix Data
If you have datasets formatted as a matrix, Kenelyze can now also directly import this data and generate network visualizations. Matrix data has been added as a new selection during the data import process:
The only prerequisite is that your file is structured as matrix: node labels should be in row 1 and column 1, while links should be indicated by a value at the intersection of nodes. There can be no value in the cell at the intersection of row 1 and column 1. Values in matrix cells are automatically added as link weights when they are numerical; if values are non-numerical, they are added as a link attribute.
Fixes and improvements
This release also fixes and improves the following:
– Fixed an issue where network statistics would not be updated with certain types of networks
– Fixed an issue where node metadata in the search bar and filters would not be updated after editing metadata in the Table View
– Color legends now show a maximum of 100 entries to avoid performance issues with attributes consisting of large numbers of unique values
– Improved rendering/scaling of fixed-size node labels on zoom
Feedback and support
If you have any feedback or questions about the new functionality, please send a message to support@kenelyze.com.
