Business Rules / Recommendation gist

Business Rule Recommendation gist

= Business Rule / Recommendation gist =
 
In this simple example, we want to highlight the power of graphs to describe, discover, visualise and implement powerful business rule-based recommendations.

In the example, we will create a simple graph containing
	- a +person+ ("Rik")
	- a +city+ ("London")
	- an +age+ ("39")
	- a +child+ ("Toon")
and all the required relationships from the person to the city, to his age, and his child. 

Then we can also immediately add an "inferred" relationship to a potential recommendation for a purchase by the person. Let's say that, based on historical data, pattern discovery, statistics or some other black magic, we know that a person living in London, age 35-40 and with a child should receive a promotional offer for an +Xbox 360+. That would be a "rule" or "recommendation" that we would want to present to the end user.

With the following query, we can create the graph:

[source,cypher]
----
// creating the based nodes and relationships

CREATE (rik:person{name:'Rik'}), (london:city{name:'London'}), (age:age{name:'39'}), (kid:child{name:'Toon'})
CREATE london<-[:LIVES_IN]-rik-[:HAS_AGE]->age
CREATE rik-[:HAS_KID]->kid

// create the "inferred" node and relationship

CREATE (xbox:console{name:'Xbox 360'})
CREATE rik-[:SHOULD_RECEIVE_PROMO_FOR]->xbox
----

Visually, this simple graph looks like this:

//graph1

Hover over the nodes to see the +name+ node property in the Graph above.

It then becomes quite easy to see that this is almost like the equivalent - but a much richer, versatile equivalent of having a traditional "business rule" that would say:

[source]
----
IF
	rik lives in london,
	AND is of age 39
	AND has a kid
THEN
	propose Xbox 360 promo
ELSE
	proceed with something else ;-) ...
----

In graph terms, we would express this as a cypher query:

[source,cypher]
----
MATCH
	rik:person-[:HAS_AGE]->age:age,
	rik:person-[:LIVES_IN]->london:city,
	rik:person-[:HAS_KID]->kid:child,
	rik:person-[:SHOULD_RECEIVE_PROMO_FOR]->sometoy
WHERE
	rik.name='Rik'
RETURN	
	rik.name, sometoy.name;
----

You can play around with this some more in the console below.

//setup
//hide
[source,cypher]
----
CREATE (rik:person{name:'Rik'}), (london:city{name:'London'}), (age:age{name:'39'}), (kid:child{name:'Toon'})
CREATE (xbox:console{name:'Xbox 360'})
CREATE london<-[:LIVES_IN]-rik-[:HAS_AGE]->age
CREATE rik-[:HAS_KID]->kid
CREATE rik-[:SHOULD_RECEIVE_PROMO_FOR]->xbox
----

//console

Hope this example is useful. Enjoy!

Hi @rvanbruggen! I've updated your gist to support the latest Neo4j 2.0 milestone: https://gist.github.com/cleishm/7311739. Can you update this version likewise?