Full Webinar: https://info.tigergraph.com/graph-gurus-25 A new weapon is available for businesses wanting to accomplish more with Hadoop: native parallel graphs can reveal the connections across multiple domains and datasets in data lakes and provide powerful insights to deliver superior outcomes. In this webinar we will explain how native parallel graphs can analyze the information in data lakes to enable the following outcomes: Recommending next best actions such as promoting a student loan to someone heading off to college, advocating life insurance to a newly married couple, and so on Improving network utilization by analyzing petabytes of data collected from millions of IoT devices across a smart grid Accelerating M&A activity by intelligently merging data lakes from multiple businesses.

1. © 20190 TigerGraph. All Rights Reserved
Graph Gurus 25
Unleash the Business Value of Your Data Lake
with Graph Analytics
1

2. © 2019 TigerGraph. All Rights Reserved
Today’s Presenter
2
David Schexnayder
Senior Sales Engineer
● 20+ years in tech industry
● 7+ years with Hadoop / Big Data
● Bachelor of Science from North Carolina State
University

3. © 2019 TigerGraph. All Rights Reserved
Some Housekeeping Items
● Although your phone is muted we do want to answer your questions -
submit your questions at any time using the Q&A tab in the menu
● The webinar is being recorded and will uploaded to our website shortly
(https://www.tigergraph.com/webinars-and-events/) and the URL will be
emailed you
● If you have issues with Zoom please contact the panelists via chat
3

4. © 2019 TigerGraph. All Rights Reserved
Our Agenda
4
Introduction1
4
3
2 Data Lakes - Characteristics and Challenges
Why Graph DB and Analytics
Why TigerGraph
5 How TigerGraph Fits
6 Resources

5. © 2019 TigerGraph. All Rights Reserved
Relational Database Key-Value Database Graph Database
Customer
XXXXXX
Product
XXXXXXXX
Supplier
XXXXXXXX
Location
XXXXXXXX
Order
XXXXXXXXX
Product
Customer
Supplier
Location
KEY VALUE
XXXXX
Order
Customer
Produc
t
Supplier
• Rigid schema
• High performance for transactions
• Poor performance for deep analytics
• Highly fluid schema/no schema
• High performance for simple transactions
• Poor performance deep analytics
• Flexible schema
• High performance for complex transactions
• High performance for deep analytics
Location 1 = Delivery Location
Location 2 = Warehouse
Location 2
Product
Payment
PURCHASED
RESIDES
SHIPSTO
PURCHASED
SHIPS FROM
A
C
C
EPTED
MAKES
The Evolution of Databases
XXXXX
XXXXX
XXXXX Location 1
N
O
TIFIES
5

6. © 2019 TigerGraph. All Rights Reserved
Characteristics of Data Lakes
Up to Petabytes of data from diverse sources
• Non-relational and relational data
• Web logs
• Social media feeds
• Collected data
Multiple engines to process and analyze data
• Spark
• Hive
• Impala
Data stored in a variety of formats
• NoSQL Databases (HBase, Cassandra)
• Columnar Storage (Parquet, ORC)
• Flat files, sequence files
6

7. © 2019 TigerGraph. All Rights Reserved
Common Components of Data Lakes
7

8. © 2019 TigerGraph. All Rights Reserved
Challenges of Data Lakes
● Data sprawl and duplication
● Gaining real-time insights very difficult
● Security and data governance
● Table joins are still expensive!
8

9. © 2019 TigerGraph. All Rights Reserved
Graph to the Rescue
TigerGraph provides the ability for businesses to put the data
in your lake to work:
• Faster insights into your connected data
• Hybrid Transactional and Analytical Processing (HTAP)
• Connect multiple data sources expressed in graph
• Distributed efficient storage and parallel processing
9

10. © 2019 TigerGraph. All Rights Reserved
Graph Can Accelerate Your Use Cases
TigerGraph can run optimized use cases to put the data in
your lake to work:
• Fraud and Anti Money Laundering
• Next-generation recommendation engines
• Customer journey / Customer 360
• Master Data Management
• Entity Resolution
• IT Optimization
10

11. © 2019 TigerGraph. All Rights Reserved
Blocking Fraud Before Payment Is Made
● 30% increase in fraud detection
● Integrated 20+ data sources,
handling 2 billion real-time
events per day, 1 trillion edges
● Distributed across 40 standard
servers (20 with 2x replication)
● Reduced working cycle from
one hour to minutes
● Business logic implemented in
10k of C++ lines
11

12. © 2019 TigerGraph. All Rights Reserved
Increase Revenues with Recommendations
Person A Person C
Shops at Store Y
Bought item X online
from Store Y
Knows/likes/follows B B commented (to C)
Person B
C (also) shops at store Y
Is located at Z
Item X is stored at
Warehouse W
Item X has
tag/feature F
Customer
Visit Click Path
Would you like…?
B likes hobby H
12

13. © 2019 TigerGraph. All Rights Reserved
Building a Customer 360° Data Hub
Business Challenge
• Combine all available data for customer with
transactions to improve business outcomes
Solution
• Build on top of current investments in master data
management, data warehouse/Hadoop data lake
and NoSQL repositories
• Find new relationships among data to drive better
fraud and money laundering detection, credit risk
scoring and monitoring, product & service
marketing, cross-sell and up-sell recommendation
for higher revenue & profits
• Analyze temporal (Time Series) and spatial data to
find new patterns and insights
• Expand schema (attributes/fields, relationships) to
accommodate new data sources & use cases
13

14. © 2019 TigerGraph. All Rights Reserved
Entity Resolution for MDM with TigerGraph
14

15. © 2019 TigerGraph. All Rights Reserved
Metadata Management with Graph
● Metadata is “Data about other Data”
● Types of Metadata
○ Structure - Organizational units, operations, IT infrastructure, stages/tates, data
collections, database schema
○ Processes - business, operational, software, data processing
○ Process flow = directional (“Directed”) graph
● Uses
○ Establish and manage policies and processes
○ Manage data lineage, integration and sharing
○ Ensure compliance (financial, GDPR, HIPAA, etc.)
○ Security, access control
○ Identity and eliminate redundancies and inefficiencies
15

16. © 2019 TigerGraph. All Rights Reserved
Improved Energy Management System
• Business Challenge: Monitor complex energy
infrastructure to detect and manage power
overloads and outages
• Solution:
• Model power system using real-time
operational Graph to accelerate state
estimation & power flow calculation
(no data preparation needed)
• Leverage massively parallel computing
in Graph for bus ordering &
admittance graph forming to balance
load spikes
• Visualize energy computation results in
Graph for contingency analysis and
action plan
• Business Value: Deliver faster than real-time
EMS (ie, EMS capable of completing
execution within a SCADA sample cycle of 5
seconds)
16

17. © 2019 TigerGraph. All Rights Reserved
Delivering Better Outcomes
Customers
Supplier
Employee
Device
IoT Signals
Orders
Payments
Shipments
Invoices
Visits
Downloads
Master Data
Operational Data
Data
Warehouse
Data Mart Data Lake NoSQL
Historical
Data
Queries / Lookups,
Comprehensive Graph Patterns
and Algorithms
Graph-Computed
Features
Batch and Streaming
17

18. © 2019 TigerGraph. All Rights Reserved
Hadoop Interoperability
Several ingestion tools to integrate Your data lake:
● Spark via JDBC
● HDFS Connector
● S3 Connector
● JDBC from RDBMS
18

19. © 2019 TigerGraph. All Rights Reserved
Testing Setup and Configuration
• Cloudera Quickstart VM 5.13.0 and example data
• TigerGraph Developer Edition 2.5.0
19

20. © 2019 TigerGraph. All Rights Reserved
Hadoop Workflow
• Use sqoop to pull transactional data from RDBMS
• Build Hive tables and store on HDFS as parquet
• Query data with Impala, Hive, Spark
20

21. © 2019 TigerGraph. All Rights Reserved
Spark and TigerGraph Data Pipeline
Static
Data
Sources
TigerGraph
JDBC
Driver
Streaming
Data
Sources

22. © 2019 TigerGraph. All Rights Reserved
JDBC Driver (1.2)
• Type 4 driver
• Support Reads and Writes bi-directional data flow to
TigerGraph
• Read: Converts ResultSet to DataFrame
• Write: Load DataFrame and files to vertex/edge in
TigerGraph
• Supports REST endpoints of built-in, compiled and
interpreted GSQL queries
• Open Source:
https://github.com/tigergraph/ecosys/tree/master/etl/tg-jdbc-driver
22

23. © 2019 TigerGraph. All Rights Reserved
JDBC Driver Installation on Cloudera
• Upgrade Java to 1.8.0 openJDK
• Convert to parcels
• Install Spark 2.4.0 parcel (optional)
• Build the TigerGraph JDBC driver (from github)
• Install the JDBC driver and add it to ‘classpath.txt’
23

24. © 2019 TigerGraph. All Rights Reserved
Relational Data Schema
24

25. © 2019 TigerGraph. All Rights Reserved
Graph Data Schema
25

26. © 2019 TigerGraph. All Rights Reserved
Hadoop —> TigerGraph Workflow
• Use Apache Spark to read the data on HDFS in parquet
format
• Use the TigerGraph Spark JDBC connector to load the data
into the graph
• Explore the graph
26

27. © 2019 TigerGraph. All Rights Reserved
Demo!
27

28. © 2019 TigerGraph. All Rights Reserved
What Makes TigerGraph Fast and Scalable?
Designed from ground up for computational parallelism
• Native graph storage
• Parallel graph computation engine
• Deep link analytics in real time
• 10X+ compression
• Distributed graph for scaleout using multiple machines
28

29. Q&A
Please submit your questions via the Q&A tab in Zoom

30. © 2019 TigerGraph. All Rights Reserved
Additional Learning Resources
Start Free at TigerGraph Cloud Today!
https://www.tigergraph.com/cloud/
Test Drive Online Demo
https://www.tigergraph.com/demo
Download the Developer Edition
https://www.tigergraph.com/download/
Guru Scripts
https://github.com/tigergraph/ecosys/tree/master/guru_scripts
Join our Developer Forum
https://groups.google.com/a/opengsql.org/forum/#!forum/gsql-users
30

31. © 2019 TigerGraph. All Rights Reserved
Coming to a City Near You
Let us know if you would like to help organize a Graph Gurus
Comes To You workshop in your city
https://info.tigergraph.com/graph-gurus-request
31

32. © 20190 TigerGraph. All Rights Reserved
Thank You