NEuler: No-code graph algorithms

In this guide, we will learn about NEuler: The Graph Data Science Playground, a graph app that helps onboarding with the Graph Data Science Library.
Please have Neo4j (version 4.0 or later) downloaded and installed to use NEuler.


What is NEuler?

220px Leonhard Euler

NEuler is a no-code UI that helps users onboard with the Neo4j Graph Data Science Library. It supports running each of the graph algorithms in the library, viewing the results, and also provides the Cypher queries to reproduce the results.

We will use the terms 'Neuler' and 'The Graph Data Science Playground' interchangeably in this guide.

Installing NEuler

NEuler is a Graph Application (Graph App) that can be installed from the Graph App Gallery that comes installed by default on the Neo4j Desktop.

We can launch the Graph Apps Gallery by clicking on the right hand side of the 'Open' menu underneath a running database:

launch graph apps gallery
Figure 1. Launching the Graph Apps Gallery (Neo4j Desktop 1.3.3)

Once we’ve done that, we’ll see the following screen.

graph apps gallery
Figure 2. The Graph Apps Gallery

We can see the Graph Data Science Playground on the first row and, if we’re using Mac or Windows, we can install it automatically by clicking on the 'Install' button.

If we’re using a Linux based Operating System we’ll need to install it manually. We can get the installation link by clicking on the get link button. Copy that link onto the clipboard and then paste it into the 'Install' section of the graph apps button sub menu:

The installation link is in case you don’t want to get it via the Graph Apps Gallery!

install gds
Figure 3. Installing NEuler, The Graph Data Science Playground

Once we’ve done that, we’re can launch NEuler. It will now appear under the menu that we get by clicking on the right side of the 'Open' button under our database, and the first time that we click on it we’ll be asked to install the dependencies:

install dependencies
Figure 4. Launching The Graph Data Science Playground

The Graph Algorithms Playground won’t work unless you install those dependencies, so make sure you click on the 'Install dependencies' button.

Once the dependencies have been installed, The Graph Algorithms Playground will launch:

gds login
Figure 5. The Graph Data Science Playground - Login Screen

The credentials for the active database are pre-populated, but we can connect NEuler to any Neo4j database that has the APOC and GDS dependencies installed. If we click on the Conenct button, the dependencies will be checked, and we’ll see the following screen:

gds select database
Figure 6. The Graph Data Science Playground - Select Database Screen

By default the 'neo4j' database is selected, but we can change the selection using the dropdown. Let’s go with the default database and press the Select database button. We’ll now see the following screen:

gds home screen
Figure 7. The Graph Data Science Playground - Home Screen

We’re now ready to start using the playground.

Loading sample datasets

Our database doesn’t contain any data at the moment, but luckily NEuler contains some sample datasets that will help us learn how to use the library. We’re going to use the Game of Thrones graph for the rest of this guide, and we can import this by following the screens in the GIF below:

gds sample dataset
Figure 8. The Graph Data Science Playground - Importing the Game of Thrones dataset

It will take about 20 seconds to get the data loaded and once its imported, the home screen will update to look like this:

gds home with data
Figure 9. The Graph Data Science Playground - Home Screen

Now that we’ve got some data loaded, it’s time to start using the algorithms.

Configuring algorithms

We’ll click on the Run Single Algorithm link on the top menu, which will take us to the following screen:

gds run single algorithm
Figure 10. Choosing an algorithm

On this screen we can select an algorithm from any of the categories, configure the projected graph that the algorithm will be run against, as well as other parameters for more complex algorithms.

The form for configuring the PageRank algorithm, one of the most well known centrality algorithms, is shown in the screenshot below:

gds select pagerank
Figure 11. Configuring the PageRank algorithm

Once we’re happy with the configuration, we can execute the algorithm by pressing the 'Run' button.

Exploring results

Results from running the algorithm will be shown on the Results tab. The output from running the PageRank algorithm against a projected graph of the interactions between Games of Thrones characters in Season 1 of the popular TV show can be seen in the diagram below:

gds pagerank results
Figure 12. Running the PageRank algorithm against Season 1 of Game of Thrones

We can see that Ned is the most important character in the show in the early days. If we want to compute the results for another season, we can do that by changing the 'Relationship Type'.

We can also view the results in chart or graph visualization formats. We can see the chart representation of PageRank run against the Game of Thrones graph in the screenshot below:

gds chart
Figure 13. A chart showing results of running the PageRank algorithm

And a graph visualization of the algorithm run in the diagram below:

gds graph viz
Figure 14. A graph visualization of running the PageRank algorithm

We can change configure the properties that select the caption, node size, and node color by using the dropdown menus above the visualization.

Seeing the code

Running algorithms in NEuler is intended as the first step in the process of learning how to use graph algorithms. The next step is learning how to run the algorithms using the underlying procedures, and NEuler helps with this by describing the procedure calls in the Code tab. The procedure call and parameter setup for running the PageRank algorithm is shown in the diagram below:

gds code view
Figure 15. The code for running the PageRank algorithm

We can use the 'Copy' button on the top right of each window to copy the code fragment to our clipboard. Alternatively, we can create a Neo4j Browser guide by clicking on the 'Send to Neo4j Browser' button.