Visualizing millions of nodes and relationships

[Raj725]

I want to store data with millions of nodes(64M) and relationships(60M+) in Neo4j and create a visualization to explore it.
I want to visualize all of them.
What is the best tool to visualize?

[david_allen]

Could you describe a bit more what you want to see? Visualization is a deep topic. If you want to put millions of things on the screen, generally this is because people want to see the overall shape of the graph. Does this describe you? Or are you seeking more to visualize just portions at a time?

Typically when people are thinking about visualization it's either exploratory (i.e. I don't know what I'm looking for and I just want to get an idea of what's out there) -- and sometimes it's specific question-oriented.

If you have a big graph, which of these are you interested in?

You might want to check out:

• Neo4j Bloom
• Neo4j Gephi

[michael_hunger]

You can also check out our visualization series: https://medium.com/neo4j/tagged/data-visualization

@William_Lyon also integrated Neo4j with Graphistry:

github.com

graphistry/pygraphistry/blob/master/demos/Neo4jTwitter.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Neo4j Twitter Trolls Tutorial\n",
    "\n",
    "**Goal**: This notebook aims to show how to use PyGraphistry to visualize data from [Neo4j](https://neo4j.com/developer/). We also show how to use [graph algorithms in Neo4j](https://neo4j.com/developer/graph-algorithms/) and use PyGraphistry to visualize the result of those algorithms.\n",
    "\n",
    "*Prerequesties:* \n",
    "* You'll need a Graphistry API key, which you can request [here](https://www.graphistry.com/api-request)\n",
    "* Neo4j. We'll be using [Neo4j Sandbox](https://neo4j.com/sandbox-v2/) (free hosted Neo4j instances pre-populated with data) for this tutorial. Specifically the \"Russian Twitter Trolls\" sandbox. You can create a Neo4j Sandbox instance [here](https://neo4j.com/sandbox-v2/)\n",
    "* Python requirements:\n",
    "  * [`neo4j-driver`](https://github.com/neo4j/neo4j-python-driver) - `pip install neo4j-driver`\n",
    "  * [`pygraphistry`](https://github.com/graphistry/pygraphistry/) - `pip install \"graphistry[all]\"` \n",
    "  "
   ]
  },
  {

This file has been truncated. show original

[Raj725]

Thanks , @david.allen, and @michael.hunger for clarification and valuable suggestions.

I have a graph with Millions of nodes and each node is connected with few nodes and have several relationship with each node.
Each relationship/edge has a property with a double value(0.0 to 1.0).
I want to filter the graph on this properties and show sub-graph(all the nodes and relationships) on the screen.

I want to apply filters something like given below on the relationship properties:

if name.levenstein <0.1& if name.jackard<0.15 & if name.soundex<0.201 & if name.math_rating==0 & if name.metaphone < 0.201 & if name.nysiis<0.113.

Thanks for your suggestions, I will check them.

[William_Lyon · Accepted solution]

For your use case it sounds like Graphistry would be a good fit. Graphistry can handle visualizing millions of nodes and then filtering based on property values / range.

The example that Michael linked above shows how to load data into Graphistry from Neo4j and then from there the Graphistry UI allows for specifying filters based on property values.

[Raj725]

Thanks @William_Lyon ️ for this help.