Introduction

CS-Students is a sample knowledge graph that can be used to model any classroom that is relation based (a classroom that has lots of different people, pairs, groups, periods; things like that). This notebook is meant to demonstrate how to query this knowledge graph based on the preset things that are currently in the sample data.

Data

Currently, the data consists of the following categories (which are labels in Neo4J):

Period (most teachers would find it useful to have their data divided into class periods)
Scrum Team (basically, the largest overarching groups of students, under period)
Pairs (pairs of students within scrum teams)
Person (object that represents each student)

The data also has the following relationships

IN_PERIOD (to relate teams to the period they are in)
CROSSOVER (a relationship that exists to connect pairs of teams which work together)
PAIR_SHARE (relationship defining students who are paired together)
IN_GROUP (to relate people to the group they are in)

Setup in Anaconda

1: Install Anaconda

2: On windows, open the Anaconda command prompt. On mac, open the normal terminal.

3: Run the following commands:

conda install jupyter

conda install py2neo

conda install pandas

conda install neo4jupyter

4: If you haven't already, then clone this repository to a location that you can navigate to easily in your files.

5: The commands above have configured your conda environment. Now, simply run the command

jupyter notebook

6: This should open up something in the browser. You will see all of your files. Navigate to the directory which has the github repository clone.

7: Then, navigate to this notebook. You should be able to run this notebook and experience all of its features now!

Setup in an IDE like VSCode

1: In the IDE terminal, run the following commands:

pip install jupyter

pip install py2neo

pip install pandas

pip install neo4jupyter

2: You're all set to run the notebook!

import pandas as pd
import getpass
from py2neo import Graph, Node, Relationship
from py2neo import GraphService
import neo4jupyter
neo4jupyter.init_notebook_mode()

# The password is made private. 
password = getpass.getpass(prompt='Password: ', stream=None) 
graph = Graph("neo4j+ssc://f9708be2.databases.neo4j.io", auth=("neo4j", password))

# database
gs = GraphService("neo4j+ssc://f9708be2.databases.neo4j.io", auth=("neo4j", password))

neo4jupyter.draw(graph, {"PERSON": "name"})

query = "MATCH(n:SCRUM_TEAM) RETURN n.name AS TeamName, n.runtimelink AS RuntimeLink"
graph.run(query).to_data_frame()
TeamName RuntimeLink
0 Team 1 www.team1.com
1 Team 2 www.team2.com 
query = "MATCH(n1:PERSON)-[rel:IN_GROUP]->(n2:SCRUM_TEAM {name: \"Team 1\"}) RETURN n1.name AS Person, rel.Job AS Job"
graph.run(query).to_data_frame()
Person Job
0 Doug Deployment Manager
1 Carl Technical Officer
2 Bob Design Manager
3 Alice Scrum Master 
query = "MATCH(n1:PERSON)-[rel:IN_GROUP]->(n2:SCRUM_TEAM {name: \"Team 2\"}) RETURN n1.name AS Person, rel.Job AS Job"
graph.run(query).to_data_frame()
Person Job
0 Henry Design Manager
1 George Technical Officer
2 Ferb Deployment Manager
3 Eve Scrum Master

The following cells show how to do CRUD operations

You can do more than just query! To see your changes in effect, rerun the visualization in the third cell.

def CreatePerson(personName):
    tx = graph.begin()
    createdPerson = Node("PERSON", name = personName)
    tx.create(createdPerson)
    graph.commit(tx)

def CreatePersonTester():
    n = input("Enter a person name: ")
    CreatePerson(n)

# Run the CreatePersonTester function to try creating a node. You should be able to find your new node!
CreatePersonTester()

# Note that this breaks with duplicate names. This is simply for demonstration purposes.
# In a real application, you need to have a way to query by unique ID
def UpdateAPerson(personName, newName):
    graph.evaluate("MATCH (n:PERSON) WHERE n.name =\"" + str(personName) + "\" SET n.name=\"" + str(newName) + "\"")

def UpdateAPersonTester():
    o = str(input("Enter the name you want to change: "))
    n = str(input("Enter the new name: "))
    UpdateAPerson(o, n)

# Try updating the node that you created in the last code block!
UpdateAPersonTester()

# Note that this breaks with duplicate names. This is simply for demonstration purposes.
def DeleteAPerson(personName):
    node = graph.evaluate("MATCH (n:PERSON) WHERE n.name =\"" + str(personName) + "\" RETURN n")
    try:
        graph.delete(node)
    except:
        return

def DeleteAPersonTester():
    x = str(input("Enter a person name: "))
    DeleteAPerson(x)

# You can try deleting the node that you created by simply entering the person name
DeleteAPersonTester()