https://github.com/dataiku/dss-plugin-graph-analytics
Visualise graph and compute graph features.
https://github.com/dataiku/dss-plugin-graph-analytics
Last synced: 7 days ago
JSON representation
Visualise graph and compute graph features.
- Host: GitHub
- URL: https://github.com/dataiku/dss-plugin-graph-analytics
- Owner: dataiku
- Created: 2020-06-05T12:51:44.000Z (almost 6 years ago)
- Default Branch: master
- Last Pushed: 2025-02-07T14:44:26.000Z (about 1 year ago)
- Last Synced: 2025-02-07T15:22:16.736Z (about 1 year ago)
- Language: Python
- Size: 248 KB
- Stars: 3
- Watchers: 21
- Forks: 1
- Open Issues: 7
-
Metadata Files:
- Readme: README.md
- Changelog: CHANGELOG.md
Awesome Lists containing this project
README
# Dataiku Graph Analytics Plugin
Version: 1.0.5
Compatible with DSS version: 12.0.0 and higher
## Overview
This plugin offers three recipes and one custom chart to create graphs and compute graph algorithms:
* **Projected graph**: for example, create a graph of Product (dataset of Product pairs) from a dataset of Product - Customer pairs where 2 products are linked if they share a same customer.
* **Graph features**: from a dataset of edges (Source - Target pairs), compute some graph algorithms to get node features such as degree, closeness centrality, pagerank, ... . It can output either a dataset of edges with the new source and target graph features in each row, or output a dataset of nodes and their corresponding new graph features.
* **Graph clustering**: from a dataset of edges, compute some graph clustering algorithms and assign nodes to their cluster Id. Same as the Graph features recipe, it can output a dataset of edges or nodes with their newly computed clusters Id.
This plugin also provides a custom chart to visualize your graph:
* You can specify the maximum number of nodes you want to display in the chart.
* You can filter nodes based on some column conditions.
* You can select some columns:
* to color nodes (nodes with same column value will have the same color)
* for node sizes (node sizes will depend on the column value)
* for edge widths and label
* When double-clicking on a node, itself and its first-degree neighbor will be highligthed.