3. ● Commercial Engineer
● 20+ years in IT of which
● 20+ years in data driven projects & analytics
● 10 years in system integrator
● 10 years in the wonderful world of startups
Personal:
1 wife, 2 kids, 1 dog, 7 drumsets
3
Introductions
Jan Aertsen
Sr Director PS EMEA
Please connect:
jan.aertsen@neo4j.com
linkedin.com/in/janaertsen/
6. Goals
“Know” our customers
through consistent engagement
Improve customer experiences
and outcomes
Effectively identify risk
and mitigate churn
Identify revenue growth
opportunities
6
Understand innovative
power of graph
Correctly and efficiently
apply graphs to business problem
Implement/adapt fast
and manage risks
Leverage software investment
1
2
3
4
11. 11
Innovation Lab
Help companies accelerate innovation
through graph thinking
How we do it
Generate and prototype graph projects
together with customers and prospects
Format
3.5 day workshop, 2-3 Neo4j participants
Outcome
Provide a deep understanding of graph
thinking and the innovation opportunities of
adapting graph technology
12. 12
Graph modeling session
Verify use case feasibility
How we do it
Interactive requirements whiteboarding and
brainstorm session
Format
0.5 day workshop, 1 expert modeler
Outcome
Validated graph modeling and graph access
scenarios. Understanding of use case
complexity
13. 13
Reference project
Talk with expert(s) (SA, PM, … ) through
implementation of a similar use case
How we do it
Discussion and presentation
Format
1-2hr phone call
Outcome
Understand high level project set-up: T-shirt
sizing of your project, identifying typical
hurdles, identify skills/team needed,
architectural components.
17. 17
Why Graph DB for MDM?
● Connect data from heterogeneous sources
○ connect data in movement
○ connect data at rest
● Agility
○ connect additional sources as needed
○ without strict schema 360 degree view is possible
○ without strict schema, alternatives can be modelled
● Lineage and traceability
○ tracking how/when/what data was loaded = graph
○ which can be stored with the data
● Intuitiveness
○ Connected data can be shared in the entire organisation
● Speed:
○ Access to MDM data with high performance 24x7 enabled
19. 1919
Bootcamp
Accelerate graph learning to allow
Customers to evaluate graph technology
How we do it
Hands on training and prototyping
Format
5 day training/workshop and Q&A
Outcome
Technical team has overall understanding of
technical capabilities of Neo4j based on hands
on experience with the toolset.
20. 2020
Proof of concept
Build a small, working solution, proving out a
select set of business requisites
How we do it
Clear scope, design and build
Format
Project
Outcome
Working solution (limited features)
Demo/presentation
Backlog and roadmap for further extensions
21. 2121
Solution design WS
Requirements analysis and solution design
exercise for full graph solution
How we do it
Requirements collection, analysis, create
product backlog and solution architecture
Format
Workshop (typically 5 days but depends … )
Outcome
Product backlog
Solution architecture/design document
Suggested project plan / road map
22. 22
PS contributions in project definition
Validated choice of Neo4j as back bone to solve “graph problem”
BUT ALSO:
• Awareness of data integration/quality challenges
• Wider application design:
• UI/UX and APIs
• Security aspects
• Operations and admin aspects
• CI/CD
• Project definition: known hurdles, risks, …
26. 2626
Solution Audit &
Upgrades
Revisit requirements, ensure alignment with
product roadmap, leverage new features
How we do it
Audit Workshop & training
Outcome
Recommendations for
1) improved solution
2) Upgrade
3) New features
27. ● Strike a balance between schemaless and enforcing constraints
● Use graph to model “uncertainty” and alternatives
● Consider data integration with data in movement or data at rest
● Picking right data integration tools
● Handling master data update priorities and business logic
● How to convince end-users the new data store is correct
27
Where we make a difference on MDM implementation
29. Add structure to unstructured data
Relate the unstructured to existing
Using taxonomies and ontologies
Use free text search
Ability to easily navigate the
unstructure
29
Knowledge graphs
Pdf Files E-Mails
TECHNICAL
DATA
Relational
DBs
3rd party
Open SourceCRM
OCR / NLP
30. 30
Types of knowledge graphs
Internal knowledge
documents & files, with
meta data tagging
External data source
aggregation mapped to
entities of interest
Context Rich Search External Event Insight
Sensing
Enterprise NLP
Graph technical terms,
acronyms, abbreviations,
misspellings, etc.
Examples:
• MDM, Search
• Customer support
• Document classification
Examples:
• Supply chain/compliance risk
• Market activity aggregation
• Sales opportunities
Examples:
• Improved search
• Chatbot implementation
• Improved classification
Context Independent
Warehouse
Real-Time WarehouseLogical Warehouse
31. 31
Where we make a difference on KG implementation
• Understand how to deal with unstructured data
• Design a data load strategy
• Working with taxonomies, ontology, context, multi-language
• Identify tools and partners to assist with
the non-Neo4j parts of the project like NLP, UI, etc
32. 32
Staffing model
Solution
Architect
Lead
Consultant
Graph
Consultant
ETL
Consultant
Lead Technical Consultant
Main technical contact and Neo4j
solution architect
Maintains architectural design
documents on Neo4j side
Guarantee holistic view across all
Neo4j interventions
Primary technical contact for
Customers
Engagement
Manager
SOW
Engagement Manager
Approve and manage scope
Assign resource of required
experience and capability
Regularly review and plan delivery
Manage change requests, resolve
blockers and point of contact for
escalations
Test / QA
Consultant
Team
Neo4j PS consultants may
contribute skills covering more
than one function
Neo4j may utilise partners who
also compliment the Neo4j
ecosystem
Neo4j can embed customer
resource into the team to
accelerate learning
UI/UX
consultant
34. 34
Network Management
Unified Network Inventory -> Network Reconciliation
Service Orchestration -> Automation
Service Assurance -> Fault Management
Planning -> Traffic Intelligence
Performance and Quality Management and
Analytics
35. Network Management projects with Neo4j
35
● Identify reusable modules
○ Connectors to standard sources
(Inventory Systems, Network Discovery Tools, Element Managers)
○ Path Computation Algos (Dijkstra, Steiner…)
○ Impact Analysis / Root Cause Analysis
○ UI: Map Based Circuit View
● Neo4j Field Engineering guides and helps you through the configuration and utilisation of the
modules
● Building integrations with your architecture elements as needed
○ Producers : Inventory Systems, Network Managers, Network Discovery Tools
○ Consumers : Event Aggregation Platforms, BI tools, Orchestration Platforms
36. Leverage the knowledge and expertise of
our field team who helped world leaders
build their graph solutions...
36
39. 39
Solutions
What?
Innovative, market-leading, Neo4j-based Business Solutions for our Enterprise
Customers and Partners
Why?
• Reduce Risk and TCO
• Rapid business value via our PS or approved solution partners
• Enables customer use cases in record timeframes
• Speeds up POC development, enables solution visualization for internal selling
• Built on extensible and customizable, solution-tuned frameworks
40. 40
Neo4j Solution Frameworks: for accelerated adoption
Risk Management (for FS)Intelligent Recommendations Network Management
Human Capital ManagementPrivacy Shield (GDPR, CCPA, …)Fraud Analysis Framework
42. 4242
Cloud Managed Service
We managed your database on your
structure
How we do it
Standardized set-up and configuration
Dedicated 24/7 Neo4j monitoring
Dedicated CMS support team
Backup/restore on request
Outcome
Fully hosted database service
44. ● Our team of trained field engineers combined with the customer’s
SMEs successfully deploy:
○ Exploratory Data Analysis
○ Apply Neo4j’s Graph Algorithms
○ Apoc + Labs + Custom Libraries
○ Python Notebooks
44
Graphs and Data Science
45. Neo4j’s field team has helped customer
Data Science teams applying Graph
Algorithms to...
Pathfinding
& Search
Centrality /
Importance
Community
Detection
Complex diverse path
computation for Service
orchestration platform with major
UK telco provider.
Detecting payment fraud for
an international electrict
utility company
Entity resolution for Global
provider of animal care
services
Clinical trial similarity for
multinational
pharmaceutical company
Similarity
47. ● Neo4j database and platform is the foundation of a successful project / implementation for:
○ Master Data Management projects
○ Knowledge graphs
○ Network management
○ …. and multiple other use cases and industries
● Neo4j Field Engineering / Services assists and enables you for successful project
implementations throughout the full cycle of
○ Grap evaluation
○ Project definition
○ Implementation
○ Hosting
… and beyond
47
Conclusions