Creating the DataFrame

First of all, let's export the data from Neo4j to a pandas DataFrame using the Neo4j Python driver (similar to the method used in Chapter 8, Using Graph-Based Features in Machine Learning):

import pandas as pd
from neo4j import GraphDatabase

driver = GraphDatabase.driver("bolt://localhost:7687", auth=("neo4j", "<YOUR_PASSWORD>"))

cypher = """
MATCH (u)
MATCH (v) 
WHERE u.id < v.id  // exclude u = v
AND NOT ( (u)-[:KNOWS_T2]-(v) )
WITH u, v,  gds.alpha.linkprediction.adamicAdar(
                u, v, {
                        relationshipQuery: "KNOWS_T1",
                        direction: "BOTH"
                }
            ) as score
RETURN  u.id as u_id, 
        v.id as v_id, 
        score, 
        EXISTS( (u)-[:KNOWS_T1]-(v) ) as label
"""

with driver.session() as session:
    rec = session.run(cypher)

df = pd.DataFrame.from_records(rec.data())

Remember to update the preceding code...