Today’s complex data is big, variably-structured and densely connected. In this session we’ll look at how size, structure and connectedness have converged to change the way we work with data. We’ll then go on to look at some of the new opportunities for creating end-user value that have emerged in a world of connected data, illustrated with graph search examples implemented using the Neo4j graph database.

1. Ian
Robinson,
Neo
Technology
Graph
Search:
The
Power
of
Connected
Data
Twi$er
@iansrobinson
#neo4j

2. Outline
• Data
complexity
• Graph
databases
–
features
and
benefits
• Querying
graph
data

3. Data
Complexity
complexity = f(size, semi-structure, connectedness)

4. Data
Complexity
complexity = f(size
, semi-structure, connectedness)

5. Semi-­‐Structure

6. Semi-­‐Structure
USER_ID
FIRST_NAME
LAST_NAME
315
Ian
Robinson
EMAIL_1
EMAIL_2
ian@neotechnology.com
iansrobinson@gmail.com
Email:
ian@neotechnology.com
Email:
iansrobinson@gmail.com
Twi$er:
@iansrobinson
Skype:
iansrobinson
0..n
USER
FACEBOOK
NULL
TWITTER
SKYPE
@iansrobinson
iansrobinson
CONTACT
CONTACT_TYPE

7. Social
Network

8. Network
Impact
Analysis

9. Route
Finding

10. Recommenda]ons

11. Logis]cs

12. Access
Control

13. Fraud
Analysis

14. Securi]es
and
Debt
Image:
orgnet.com

15. Graphs
Are
Everywhere

16. Graph
Databases
• Store
• Manage
• Query
data

17. Neo4j
is
a
Graph
Database

18. Labeled
Property
Graph

19. The
Power
of
Rela]onships
Connectedness
• Pre-­‐computed
joins
• Several
million
joins
per
second
per
thread
per
core
Variable
Structure
• Defined
with
regard
to
node
instances,
not
classes
of
nodes
– Contrast
with
rela]onal
schemas,
where
foreign
key
rela]onships
apply
to
all
rows
in
a
table

20. Graph
Database
Benefits
“Minutes
to
milliseconds”
performance
• Millions
of
‘joins’
per
second
• Consistent
query
]mes
as
dataset
grows
Fit
for
the
domain
• Lots
of
join
tables?
Connectedness
• Lots
of
sparse
tables?
Semi-­‐structure
Business
responsiveness
• Easy
to
evolve

21. Querying
Graph
Data
• Describing
graphs
• Crea]ng
nodes,
rela]onships
and
proper]es
• Querying
graphs

22. How
to
Describe
a
Graph?

23. Cypher
Padern
(rest)<-[:HAS_SKILL]-(ben)-[:HAS_SKILL]->(neo4j),
(ben)-[:WORKS_FOR]->(acme)

24. Create
Some
Data
CREATE (ben:Person { name:'Ben' }),
(acme:Company { name:'Acme' }),
(rest:Skill { name:'REST' }),
(neo4j:Skill= { name:'Neo4j' }),
(ben)-[:WORKS_FOR]->(acme),
(ben)-[:HAS_SKILL]->(rest),
(ben)-[:HAS_SKILL]->(graphs)
RETURN ben

25. Create
Nodes
CREATE (ben:Person { name:'Ben' }),
(acme:Company { name:'Acme' }),
(rest:Skill { name:'REST' }),
(neo4j:Skill= { name:'Neo4j' }),
(ben)-[:WORKS_FOR]->(acme),
(ben)-[:HAS_SKILL]->(rest),
(ben)-[:HAS_SKILL]->(graphs)
Iden]fier
Proper]es
RETURN ben
Label

26. Create
Rela]onships
CREATE (ben:Person { name:'Ben' }),
Iden]fier
Iden]fier
(acme:Company { name:'Acme' }),
Rela]onship
(rest:Skill { name:'REST' }),
(neo4j:Skill= { name:'Neo4j' }),
(ben)-[:WORKS_FOR]->(acme),
(ben)-[:HAS_SKILL]->(rest),
(ben)-[:HAS_SKILL]->(graphs)
RETURN ben

27. Return
Node
CREATE (ben:Person { name:'Ben' }),
(acme:Company { name:'Acme' }),
(rest:Skill { name:'REST' }),
(neo4j:Skill= { name:'Neo4j' }),
(ben)-[:WORKS_FOR]->(acme),
(ben)-[:HAS_SKILL]->(rest),
(ben)-[:HAS_SKILL]->(graphs)
RETURN ben

29. Querying
a
Graph
Graph
Local
• Find
one
or
more
start
nodes
• Explore
surrounding
graph
• Millions
of
hops
per
second

30. Which
people,
who
work
for
the
same
company
as
me,
share
my
skills?

31. Cypher
Padern
(company)<-[:WORKS_FOR]-(me)-[:HAS_SKILL]->(skill),
(company)<-[:WORKS_FOR]-(colleague)-[:HAS_SKILL]->(skill)

32. Cypher
Query
Which
people,
who
work
for
the
same
company
as
me,
have
similar
skills
to
me?
MATCH (company)<-[:WORKS_FOR]-(me:Person)-[:HAS_SKILL]->(skill),
(company)<-[:WORKS_FOR]-(colleague)-[:HAS_SKILL]->(skill)
WHERE me.name = 'ian'
RETURN colleague.name AS name,
count(skill) AS score,
collect(skill.name) AS skills
ORDER BY score DESC

33. Graph
Padern
Which
people,
who
work
for
the
same
company
as
me,
have
similar
skills
to
me?
MATCH (company)<-[:WORKS_FOR]-(me:Person)-[:HAS_SKILL]->(skill),
(company)<-[:WORKS_FOR]-(colleague)-[:HAS_SKILL]->(skill)
WHERE me.name = 'ian'
RETURN colleague.name AS name,
count(skill) AS score,
collect(skill.name) AS skills
ORDER BY score DESC

34. Anchor
Padern
in
Graph
Which
people,
who
work
for
the
same
company
as
me,
have
similar
skills
to
me?
MATCH (company)<-[:WORKS_FOR]-(me:Person)-[:HAS_SKILL]->(skill),
(company)<-[:WORKS_FOR]-(colleague)-[:HAS_SKILL]->(skill)
WHERE me.name = 'ian'
RETURN colleague.name AS name,
count(skill) AS score,
collect(skill.name) AS skills
ORDER BY score DESC
Search
nodes
labelled
‘Person’,
matching
on
‘name’
property

35. Create
Results
Which
people,
who
work
for
the
same
company
as
me,
have
similar
skills
to
me?
MATCH (company)<-[:WORKS_FOR]-(me:Person)-[:HAS_SKILL]->(skill),
(company)<-[:WORKS_FOR]-(colleague)-[:HAS_SKILL]->(skill)
WHERE me.name = 'ian'
RETURN colleague.name AS name,
count(skill) AS score,
collect(skill.name) AS skills
ORDER BY score DESC

37. Results
+--------------------------------------+
| name
| score | skills
|
+--------------------------------------+
| "Ben"
| 2
| ["Neo4j","REST"] |
| "Charlie" | 1
| ["Neo4j"]
|
+--------------------------------------+
2 rows

38. Case
Studies

39. Network
Impact
Analysis
• Which
parts
of
network
does
a
customer
depend
on?
• Who
will
be
affected
if
we
replace
a
network
element?

40. Asset
Management
&
Access
Control
• Which
assets
can
an
admin
control?
• Who
can
change
my
subscrip]on?

41. Logis]cs
• What’s
the
quickest
delivery
route
for
this
parcel?

42. Social
Network
&
Recommenda]ons
• Which
assets
can
I
access?
• Who
shares
my
interests?

43. graphdatabases.com
ts gy
en lo
i m no
pl h
m ec
Co eo T
N
of
Graph
h
Databases
Ian Robinson,
Jim Webber & Emil Eifrem
Thank
you
@ianSrobinson
ian@neotechnology.com