2. Agenda
â˘âŻ Background of graph database
â˘âŻ Areas of application
â˘âŻ Graph Intelligenceâs graph database system
3. The Background
â˘âŻ A hyper-connected world generating huge amount of
complex, inter-connected data demands a solution to
â˘âŻ Store and manage complex inter-connected data
â˘âŻ Make sense of that data
â˘âŻ Evolve with the data
â˘âŻ âGraph analysis is possibly the single most effective
competitive differentiator for organizations pursuing data-
driven operations and decisions after the design of data
capture.â - Gartner 2014
4. What is Graph Database
â˘âŻ Graph Database is a software system used to persist and
process graphs, i.e. data in terms of entities and the
relationships between entities.
â˘âŻ Nodes: represent entities such as people, businesses, accounts,
events, policies, etc.
â˘âŻ Edges: the (un)directed lines connecting nodes, represent the
relationship in between.
â˘âŻ Properties: pertinent information that relates to nodes and edges.
5. Why Graph DB?
â˘âŻ Good for problem domains that have:
â˘âŻ Innate network structures with need for pattern matching and recursive
graph search (e.g. 1000s of objects with many-many relationships, Complex
sequences and workflows)
â˘âŻ (source: Neo4j)
â˘âŻ Changes on structure of data on a regular basis
â˘âŻ Native to problem domain, and hence less modeling and programming
efforts with higher productivity
8. Network Analysis
â˘âŻ Social Network Analysis
â˘âŻ How viral are my marketing content?
â˘âŻ Who are the most influential customers?
â˘âŻ How does my company relate to other industry players?
â˘âŻ Product/Customer Community Detection
â˘âŻ Uncover different categorization of products
â˘âŻ Uncover hidden patterns of association within and cross products
groups.
9. Recommendation
â˘âŻ Cross-channel real-time targeted recommendations
based on each userâs latest activities (context)
â˘âŻ More likely to influence user behavior
â˘âŻ Reduced computational resources for batch processing
10. Information in Context
â˘âŻ Real-time Access to Cross-domain Relationships
â˘âŻ Augment traditional product information with dynamic feedback
from web and social media
â˘âŻ 360-degree view of customer across product lines, locations,
organizations and interaction channels (sales, billing, support,
mobile, web, social media, etc.)
â˘âŻ Intelligence, Crime
â˘âŻ Master data management
11. Identity and Access Management
â˘âŻ Leveraging real world relationships between people,
assets, roles, organizations and security policies.
â˘âŻ Determine authorization by tracing from individuals, through
groups, roles and products without the mismatch of traditional
hierarchical DBMS.
â˘âŻ Easily manage dynamic group membership and inter-relationships
â˘âŻ Easy to check consistency in policy updates and avoid conflict
â˘âŻ Real-time queries that are multi-dimensional and across-
hierarchies.
â˘âŻ Graph models make it easy to evolve your identity and access
management models.
12. Some Other Areas
â˘âŻ Bioinformatics:
â˘âŻ Biological DB with 111 ontologies and 50,0000+ classes/Types:
how to store and manage relationship of all these different classes/
types in RMDB?
â˘âŻ Workflow optimization
â˘âŻ Supply chain management: optimizing workflows
â˘âŻ Geo-Routing: Given the polluted segment S1, find all the upstream
segments within 50 miles of City1200.
13. Graph Intelligence
â˘âŻ Scalable graph database for real-time analytics!
â˘âŻ Highly scalable >>Neo4j
â˘âŻ Strong transactional support >OrientDB & Titan
â˘âŻ Optimized for real time dynamic graphs with snapshot
isolation: the only graph database that natively tracks evolution
of graph
â˘âŻ Schema-free modeling: SQL-compliant
â˘âŻ Fast traversals with native graph structure
â˘âŻ The only graph database innate to Hadoop ecosystem