Knowledge Graphs

for a Connected World
March 24, 2016
Benjamin Nussbaum 

@bennussbaum
www.graphgrid.com | www.atomrain.com
Introduction
Benjamin Nussbaum
20 years of Technology Innovation.

Software architecture | Database design | Server infrastructure

President & CTO of AtomRain,

one of the world’s leading NEO4J Solution Partners 

and makers of GraphGrid.
a platform by
Today’s Meetup Agenda
Knowledge Graphs for a Connected World
• What is driving the adoption of graphs
Graph Basics for AI Champions
• Where a graph fits within a web 3.0 strategy
• Why a graph is the first step to AI
• How a graph works
Graph Development for Innovation Teams
• Who does what
Graphs in Action
• Popular use cases
• Putting it all together
Q&A
A Web of Things

Generating a Web of Data
Knowledge Graphs
are driving
strategies for
Web 3.0,
The Semantic Web
A Web of Things

Generating a Web of Data
Dynamic Data
At Web Scale
The Entertainment Graph TM

560 million nodes 

1.8 billion relationships

3.0 billion properties
Continuous ingestion from dozens of
external such as Wikipedia, Netflix, Amazon
and iTunes for personalized recommendations
and social discovery of content.
The World’s Leading Graph Database
Brands and Ventures

now have access to graph platform services
Solution Partner
Brands and Ventures

now have access to graph development partners
Solution Partner
Today’s Meetup Agenda
Knowledge Graphs for a Connected World
• What is driving the adoption of graphs
Graph Basics for AI Champions
• Where a graph fits within a web 3.0 strategy
• Why a graph is the first step to AI
• How a graph works
Graph Development for Innovation Teams
• Who does what
Graphs in Action
• Popular use cases
• Putting it all together
Q&A
Knowledge

Graph
Big Data

Ingestion
Real-Time
Queries & Algorithms
Pre-Computed
Queries & Algorithm
Discovery

+ Reasoning
Personalizing Apps Smart Places Interacting Machines
Your Graph is a Data Service to “Smart” Touchpoints
DATA PLATFORM API
Data 

Science
Artificial

Intelligence
Relating

with Interaction
Acting

with Processing Layers
Serving

with Graphs
Discerning

with Patterns
Identify Link Prescribe Do ThinkPredict Sense Adapt
Apps 

access and update

the graph
Real-Time

data about customers

things, and relationships.
Algorithms 

reason

over the graph
Patterns

for best, worst, and next

steps or things.
Smart Things

send machine results 

to the graph
History 

of a machine’s

action and results.
AIs

access customer
insight

in the graph
Prediction

of a customer’s

next need or want.
A Graph Manages 

your Brand’s Evolving Knowledge
Knowledge Graph
A Graph Records a New Kind of Data
Semantic Web

and 

Knowledge Graphs
Enterprises Systems

and 

Business Transactions
For Business Operations
▪ Business Systems

generate data. 

▪ Data about Business

Orders, purchases, invoices,

customer interactions…

▪ Static System of Record 

Standard data; relationships are
not first class citizens.

CRM

System
Product
Catalog
Invoice

System
▪ Connected Customers & Smart
Things

generate data.

▪ Data about Real-World 

Concepts, people, places, things,

and their relationships.

▪ Dynamic Graph of Relationships

Discovers and learns through
patterns as relationships change

For Connected Experiences
A “Node”

in the graph
Hotel
Room
Person
A Graph Models

Real-World People, Places, and Things
Solution Partner
A “Label”

in the graph
A “Relationship”

in the graph
PREFERS
Hotel
Room
Person
HAS_AVAILABLE
A Graph Models

Contextual Relationships
Solution Partner
PREFERS
Hotel
Room
Person
“Properties”

in the graph
lastStayed: 2-10-2015
name: Hilton Hotel
name: Jane Smith
number: 315
HAS_AVAILABLE
A Graph Stores and Updates Data

about Each Thing and its Relationships
Solution Partner
PREFERS
Hotel
Room
Person
Queries

traverse the graph to discover

relevant resources
For Jane’s preferred hotel

and travel destination, 

identify available rooms,

present information to her app. HAS_AVAILABLE
Algorithms

calculate to

solve problems
- Spot Patterns.

- Prescribe Best Solution.

- Predict Results.
Queries and Algorithms

Reason over the Graph
Solution Partner
Graph Queries

Start with one “entity” and traverse the graph 

to discover linked people, places, or things
Query for a Graph
MATCH (boss)-[:MANAGES*0..3]->(sub),
(sub)-[:MANAGES*1..3]->(report)
WHERE boss.name = “John Doe”
RETURN sub.name AS Subordinate, 

count(report) AS Total
NEO4J Cypher Language
“Complex Join” in SQL
Solution Partner
Example: 

Calculates the shortest path—the least number of nodes, relationships—between two nodes
Traversal Algorithms

Navigate the graph and calculate to spot patterns or solve problems
Solution Partner
Today’s Meetup Agenda
Knowledge Graphs for a Connected World
• What is driving the adoption of graphs
Graph Basics for AI Champions
• Where a graph fits within a web 3.0 strategy
• Why a graph is the first step to AI
• How a graph works
Graph Development for Innovation Teams
• Who does what
Graphs in Action
• Popular use cases
• Putting it all together
Q&A
Subject Matter Experts work with Graph Experts

to create the conceptual model
Graph Software Engineers

create the software solution to transform and load data into the graph model
Multidisciplinary Team

ensures the quality of queries and algorithms
User Results
Ongoing Lab:
• Subject Matter Experts
(i.e Marketing)
• Data Engineer
• Algorithm Developer
Knowledge

Graph
Big Data

Ingestion
Real-Time
Queries & Algorithms
Pre-Computed
Queries & Algorithm
Discovery

+ Reasoning
DATA PLATFORM
Platform Experts

manage the scaling platform
Top Challenges

1. Query Performance
2. Algorithm Performance

3. Graph Operation

at Scale
4. Server Infrastructure 

at Scale

5. Ingestion Engines
6. Entity Resolution
API
An enterprise-grade, internet scale

data management platform
Today’s Meetup Agenda
Knowledge Graphs for a Connected World
• What is driving the adoption of graphs
Graph Basics for AI Champions
• Where a graph fits within a web 3.0 strategy
• Why a graph is the first step to AI
• How a graph works
Graph Development for Innovation Teams
• Who does what
Graphs in Action
• Popular use cases
• Putting it all together
Q&A
Master Data Management
For customer interests, product lines, store locations, org charts…
For white papers, visit neo4j.com/use-cases/
Identify & Access Management
Validates who you are, what group you belong to, and what you’re permitted to
do.
For white papers, visit neo4j.com/use-cases/
Graph Based Search
Delivers a structured result: such as a song, music attributes, artist, album, and
playlists.
For white papers, visit neo4j.com/use-cases/
Real time Recommendations
Based on past purchases, recent browsing, or friends’ purchases.
For white papers, visit neo4j.com/use-cases/
Social Network
Family, friend and follower relationships

reveal influencers, peer groups, and patterns of social behavior.
For white papers, visit neo4j.com/use-cases/
Fraud Detection
Uncovers fraud rings and patterns of unusual customer behavior.
For white papers, visit neo4j.com/use-cases/
Putting it all together: A Connected Fitness Venture
PERSONA

GOALS AND PREFERENCES
• Skill Level
• Health Conditions
• Workout Goals
• Eating Goals
• Muscle Groups
• Body Areas
• Workout Types
• Supplement Needs
CONSUMER WANTS
1. What fitness programs are best to help me accomplish my workout goals?
2. Which nutritional supplements will help me achieve my eating and workout goals?
3. Who in the community can I work out with and which workout would be good to do
together?
Scoring Algorithm 

considers importances 

the user places 

on each item
For Complete Review with Sample Queries

http://neo4j.com/graphgist/95f4f165-0172-4b3d-981b-edcbab2e0a4b
BRAND’s 

WEB OF EVERYTHING
• Supplement lines
• Fitness programs
• Social network
Putting it all together: A Connected Fitness Venture
For Complete Review with Sample Queries

http://neo4j.com/graphgist/95f4f165-0172-4b3d-981b-edcbab2e0a4b
Product Cross-Selling

aligned to users’ 

personal goals—and results
Putting it all together: A Connected Fitness Venture
For Complete Review with Sample Queries

http://neo4j.com/graphgist/95f4f165-0172-4b3d-981b-edcbab2e0a4b
Today’s Meetup Agenda
Knowledge Graphs for a Connected World
• What is driving the adoption of graphs
Graph Basics for AI Champions
• Where a graph fits within a web 3.0 strategy
• Why a graph is the first step to AI
• How a graph works
Graph Development for Innovation Teams
• Who does what
Graphs in Action
• Popular use cases
• Putting it all together
Q&A
Q&A
What do you think about Graphs?
Thank You!
Knowledge Graphs for a Connected World
March 24, 2016
Benjamin Nussbaum 

@bennussbaum
www.graphgrid.com | www.atomrain.com

Knowledge Graphs for a Connected World - AI, Deep & Machine Learning Meetup

  • 1.
    Knowledge Graphs
 for aConnected World March 24, 2016 Benjamin Nussbaum 
 @bennussbaum www.graphgrid.com | www.atomrain.com
  • 2.
    Introduction Benjamin Nussbaum 20 yearsof Technology Innovation.
 Software architecture | Database design | Server infrastructure
 President & CTO of AtomRain,
 one of the world’s leading NEO4J Solution Partners 
 and makers of GraphGrid. a platform by
  • 3.
    Today’s Meetup Agenda KnowledgeGraphs for a Connected World • What is driving the adoption of graphs Graph Basics for AI Champions • Where a graph fits within a web 3.0 strategy • Why a graph is the first step to AI • How a graph works Graph Development for Innovation Teams • Who does what Graphs in Action • Popular use cases • Putting it all together Q&A
  • 4.
    A Web ofThings
 Generating a Web of Data Knowledge Graphs are driving strategies for Web 3.0, The Semantic Web
  • 5.
    A Web ofThings
 Generating a Web of Data Dynamic Data At Web Scale The Entertainment Graph TM
 560 million nodes 
 1.8 billion relationships
 3.0 billion properties Continuous ingestion from dozens of external such as Wikipedia, Netflix, Amazon and iTunes for personalized recommendations and social discovery of content. The World’s Leading Graph Database
  • 6.
    Brands and Ventures
 nowhave access to graph platform services Solution Partner
  • 7.
    Brands and Ventures
 nowhave access to graph development partners Solution Partner
  • 8.
    Today’s Meetup Agenda KnowledgeGraphs for a Connected World • What is driving the adoption of graphs Graph Basics for AI Champions • Where a graph fits within a web 3.0 strategy • Why a graph is the first step to AI • How a graph works Graph Development for Innovation Teams • Who does what Graphs in Action • Popular use cases • Putting it all together Q&A
  • 9.
    Knowledge
 Graph Big Data
 Ingestion Real-Time Queries &Algorithms Pre-Computed Queries & Algorithm Discovery
 + Reasoning Personalizing Apps Smart Places Interacting Machines Your Graph is a Data Service to “Smart” Touchpoints DATA PLATFORM API
  • 10.
    Data 
 Science Artificial
 Intelligence Relating
 with Interaction Acting
 withProcessing Layers Serving
 with Graphs Discerning
 with Patterns Identify Link Prescribe Do ThinkPredict Sense Adapt Apps 
 access and update
 the graph Real-Time
 data about customers
 things, and relationships. Algorithms 
 reason
 over the graph Patterns
 for best, worst, and next
 steps or things. Smart Things
 send machine results 
 to the graph History 
 of a machine’s
 action and results. AIs
 access customer insight
 in the graph Prediction
 of a customer’s
 next need or want. A Graph Manages 
 your Brand’s Evolving Knowledge Knowledge Graph
  • 11.
    A Graph Recordsa New Kind of Data Semantic Web
 and 
 Knowledge Graphs Enterprises Systems
 and 
 Business Transactions For Business Operations ▪ Business Systems
 generate data. 
 ▪ Data about Business
 Orders, purchases, invoices,
 customer interactions…
 ▪ Static System of Record 
 Standard data; relationships are not first class citizens.
 CRM
 System Product Catalog Invoice
 System ▪ Connected Customers & Smart Things
 generate data.
 ▪ Data about Real-World 
 Concepts, people, places, things,
 and their relationships.
 ▪ Dynamic Graph of Relationships
 Discovers and learns through patterns as relationships change
 For Connected Experiences
  • 12.
    A “Node”
 in thegraph Hotel Room Person A Graph Models
 Real-World People, Places, and Things Solution Partner A “Label”
 in the graph
  • 13.
    A “Relationship”
 in thegraph PREFERS Hotel Room Person HAS_AVAILABLE A Graph Models
 Contextual Relationships Solution Partner
  • 14.
    PREFERS Hotel Room Person “Properties”
 in the graph lastStayed:2-10-2015 name: Hilton Hotel name: Jane Smith number: 315 HAS_AVAILABLE A Graph Stores and Updates Data
 about Each Thing and its Relationships Solution Partner
  • 15.
    PREFERS Hotel Room Person Queries
 traverse the graphto discover
 relevant resources For Jane’s preferred hotel
 and travel destination, 
 identify available rooms,
 present information to her app. HAS_AVAILABLE Algorithms
 calculate to
 solve problems - Spot Patterns.
 - Prescribe Best Solution.
 - Predict Results. Queries and Algorithms
 Reason over the Graph Solution Partner
  • 16.
    Graph Queries
 Start withone “entity” and traverse the graph 
 to discover linked people, places, or things Query for a Graph MATCH (boss)-[:MANAGES*0..3]->(sub), (sub)-[:MANAGES*1..3]->(report) WHERE boss.name = “John Doe” RETURN sub.name AS Subordinate, 
 count(report) AS Total NEO4J Cypher Language “Complex Join” in SQL Solution Partner
  • 17.
    Example: 
 Calculates theshortest path—the least number of nodes, relationships—between two nodes Traversal Algorithms
 Navigate the graph and calculate to spot patterns or solve problems Solution Partner
  • 18.
    Today’s Meetup Agenda KnowledgeGraphs for a Connected World • What is driving the adoption of graphs Graph Basics for AI Champions • Where a graph fits within a web 3.0 strategy • Why a graph is the first step to AI • How a graph works Graph Development for Innovation Teams • Who does what Graphs in Action • Popular use cases • Putting it all together Q&A
  • 19.
    Subject Matter Expertswork with Graph Experts
 to create the conceptual model
  • 20.
    Graph Software Engineers
 createthe software solution to transform and load data into the graph model
  • 21.
    Multidisciplinary Team
 ensures thequality of queries and algorithms User Results Ongoing Lab: • Subject Matter Experts (i.e Marketing) • Data Engineer • Algorithm Developer
  • 22.
    Knowledge
 Graph Big Data
 Ingestion Real-Time Queries &Algorithms Pre-Computed Queries & Algorithm Discovery
 + Reasoning DATA PLATFORM Platform Experts
 manage the scaling platform Top Challenges
 1. Query Performance 2. Algorithm Performance
 3. Graph Operation
 at Scale 4. Server Infrastructure 
 at Scale
 5. Ingestion Engines 6. Entity Resolution API An enterprise-grade, internet scale
 data management platform
  • 23.
    Today’s Meetup Agenda KnowledgeGraphs for a Connected World • What is driving the adoption of graphs Graph Basics for AI Champions • Where a graph fits within a web 3.0 strategy • Why a graph is the first step to AI • How a graph works Graph Development for Innovation Teams • Who does what Graphs in Action • Popular use cases • Putting it all together Q&A
  • 24.
    Master Data Management Forcustomer interests, product lines, store locations, org charts… For white papers, visit neo4j.com/use-cases/
  • 25.
    Identify & AccessManagement Validates who you are, what group you belong to, and what you’re permitted to do. For white papers, visit neo4j.com/use-cases/
  • 26.
    Graph Based Search Deliversa structured result: such as a song, music attributes, artist, album, and playlists. For white papers, visit neo4j.com/use-cases/
  • 27.
    Real time Recommendations Basedon past purchases, recent browsing, or friends’ purchases. For white papers, visit neo4j.com/use-cases/
  • 28.
    Social Network Family, friendand follower relationships
 reveal influencers, peer groups, and patterns of social behavior. For white papers, visit neo4j.com/use-cases/
  • 29.
    Fraud Detection Uncovers fraudrings and patterns of unusual customer behavior. For white papers, visit neo4j.com/use-cases/
  • 30.
    Putting it alltogether: A Connected Fitness Venture PERSONA
 GOALS AND PREFERENCES • Skill Level • Health Conditions • Workout Goals • Eating Goals • Muscle Groups • Body Areas • Workout Types • Supplement Needs CONSUMER WANTS 1. What fitness programs are best to help me accomplish my workout goals? 2. Which nutritional supplements will help me achieve my eating and workout goals? 3. Who in the community can I work out with and which workout would be good to do together? Scoring Algorithm 
 considers importances 
 the user places 
 on each item For Complete Review with Sample Queries
 http://neo4j.com/graphgist/95f4f165-0172-4b3d-981b-edcbab2e0a4b
  • 31.
    BRAND’s 
 WEB OFEVERYTHING • Supplement lines • Fitness programs • Social network Putting it all together: A Connected Fitness Venture For Complete Review with Sample Queries
 http://neo4j.com/graphgist/95f4f165-0172-4b3d-981b-edcbab2e0a4b
  • 32.
    Product Cross-Selling
 aligned tousers’ 
 personal goals—and results Putting it all together: A Connected Fitness Venture For Complete Review with Sample Queries
 http://neo4j.com/graphgist/95f4f165-0172-4b3d-981b-edcbab2e0a4b
  • 33.
    Today’s Meetup Agenda KnowledgeGraphs for a Connected World • What is driving the adoption of graphs Graph Basics for AI Champions • Where a graph fits within a web 3.0 strategy • Why a graph is the first step to AI • How a graph works Graph Development for Innovation Teams • Who does what Graphs in Action • Popular use cases • Putting it all together Q&A
  • 34.
    Q&A What do youthink about Graphs?
  • 35.
    Thank You! Knowledge Graphsfor a Connected World March 24, 2016 Benjamin Nussbaum 
 @bennussbaum www.graphgrid.com | www.atomrain.com