Arango logo
arango.ai 

Graph Analytics Quick Start

Reveal the structure hidden in your connected data - rank influential entities, detect communities, and spot fraud

Prerequisites

  • Access to an Arango Contextual Data Platform deployment.
  • An existing graph in ArangoDB (node and edge collections).
  • An API key or authentication token (for API access).

Run PageRank

Start an ephemeral engine

In the Running Engines panel, click Start New Engine, choose an engine size, and click Start Engine. The engine is the dedicated compute resource your algorithms run on.

Load your graph

In the Select Graph dropdown, click Load New Graph, pick your graph from the database, and click Load Graph.

Run PageRank

In Algorithm Execution, select your engine and loaded graph, choose PageRank, set parameters such as damping_factor and maximum_supersteps, and click Run Algorithm.

Monitor the job

Click View Jobs History to track the execution. Wait until the Status column of your algorithm job shows COMPLETED.

Store the results

In the Actions column of the completed job, click the icon, specify the target collection and the attribute name to store the results under, and confirm.

Start an ephemeral engine

POST https://<EXTERNAL_ENDPOINT>:8529/_platform/acp/v1/graphanalytics

Send an empty body. The response returns a serviceId whose postfix (serviceIdPostfix) you use in the engine URLs of the following calls.

Load your graph

POST https://<EXTERNAL_ENDPOINT>:8529/gral/:serviceIdPostfix/v1/loaddata 
{ "database": "_system", "graph_name": "yourGraphName" }

Returns a job_id and a graph_id.

Run PageRank

POST https://<EXTERNAL_ENDPOINT>:8529/gral/:serviceIdPostfix/v1/pagerank 
{ "graph_id": "234", "damping_factor": 0.85, "maximum_supersteps": 500 }

Returns a job_id.

Check the job progress

GET https://<EXTERNAL_ENDPOINT>:8529/gral/:serviceIdPostfix/v1/jobs/:JOB_ID

Poll the job until its progress is equal to its total. The algorithm needs to finish before you can store its results.

Store the results

POST https://<EXTERNAL_ENDPOINT>:8529/gral/:serviceIdPostfix/v1/storeresults 
{
  "database": "_system",
  "target_collection": "results",
  "job_ids": ["<job_id>"],
  "attribute_names": ["pagerank_score"]
}
You now have your nodes ranked by PageRank score, persisted to a dedicated ArangoDB collection - your most influential nodes are at the top.

PageRank is just one example. Graph Analytics supports a range of algorithms that you can combine to answer different questions about your data:

Next steps

On this page