Your first graph

Question: which papers in a small reading list are cited by other papers?

Basic · No programming knowledge needed to understand this example.

A graph records items and their connections. Here, each node is a paper; a relationship says that one paper cites another. A query is a question we ask about the recorded connections. This is a graph of relationships, not a bar chart.

Our three fictional papers have these connections:

Paper	Cites
Survey	Methods and Replication
Replication	Methods
Methods	No papers in this example

Count how often each title appears in the right column: Methods appears twice, Replication once, and Survey never. Those are the answers we expect from GraphForge. A citation count describes these records; it does not establish a paper’s quality.

Have GraphForge check the answer

Use Work with an agent to have a coding assistant run this example. You should be able to compare the returned answer with the table above without understanding Python. Initial computer setup may need help from a technical colleague or computer support person. A chat window that cannot run code cannot perform this check.

If you already know basic Python and terminal commands, follow Installation, then run the code below in a script or notebook.

Python instructions for an agent, helper, or Advanced reader

This v0.6.0 example runs in memory: its data lasts only while the program is open. It needs no account or saved project folder.

Create connected data

from graphforge import GraphForge

forge = GraphForge()
survey = forge.add_node("Paper", title="Survey")
methods = forge.add_node("Paper", title="Methods")
replication = forge.add_node("Paper", title="Replication")
forge.add_edge(survey, "CITES", methods)
forge.add_edge(survey, "CITES", replication)
forge.add_edge(replication, "CITES", methods)

These are synthetic papers. The graph says that Survey cites two papers and Replication cites Methods. add_node returns a handle you can pass directly to add_edge; you do not need to construct identities yourself.

Ask a question

result = forge.execute("""
    MATCH (:Paper)-[:CITES]->(paper:Paper)
    RETURN paper.title AS title, count(*) AS citations
    ORDER BY citations DESC, title
""")
print(result.to_pylist())

Expected output:

[{'title': 'Methods', 'citations': 2}, {'title': 'Replication', 'citations': 1}]

Methods has two incoming citations in this graph. Survey has none, so it is absent from this query’s result. The answer describes the example data, not the papers’ real-world quality or influence.

The result is an Arrow table. to_pylist() makes it easy to inspect without installing pandas or learning Arrow’s storage format.

Close the example

forge.close()

Closing this in-memory instance discards its graph. Re-running the creation cell against the same open instance adds another set of nodes; start with a new instance when repeating the example.

Check the result

The returned answer should contain Methods: 2 and Replication: 1. Survey is absent because the question asks for papers that have incoming citations. No result here means no matching connection in these records; it does not mean that nobody anywhere has cited Survey.

This example keeps its graph in memory, so closing the program discards it. Keep the example instructions to repeat it. Saving a project is a separate step.

You have completed basic graph use. To connect numerical and interview evidence, continue with Your first mixed-methods project. You can stop at either lesson; neither requires learning every GraphForge feature.

To write your own code, move to Advanced. It assumes basic Python, database, and terminal skills and explains how to save, query, and analyze a graph.