The online seminar on how to help data scientists improve data insight was held on June 8

Figure calculation | Graph database | Figure data analysis

Figure data | Knowledge map | Graph data science

Neo4j Graph data science （GDS） It aims to enable data scientists to easily achieve more accurate prediction through comprehensive graph analysis technology , Through the graph algorithm library 、 Machine learning and data science methods improve models .Neo4j Graph data science （GDS） It has been widely adopted and implemented on a large scale , Easily handle hundreds of billions of nodes and relationships . Accessible 60 Multiple powerful 、 Scalable algorithms and supervised machine learning , Use data relationships to make better predictions . from Neo4j Online live broadcast -《 Graph data science （GDS） How to help data scientists improve data insight 》 Will be in 2022 year 6 month 8 Japan 16:00 hold , Welcome to scan the code to sign up for the conference , Bring you a different graph data science .

With the continuous progress of human society , The relationship between data becomes more and more important . As a bearer of data resources “ Containers ” And the graph database that can provide query and analysis ability becomes more and more important . At present , Figure database is becoming an emerging hot field for developed countries to compete in the database field , And has formed a preliminary market scale , It is developing rapidly . It is predicted that , from 2020 - 2026 year , The size of the global map database market will be 28.6% The annual growth rate of . Although in the era of traditional database , Foreign enterprises have been occupying the absolute share of the domestic database market , But in the age of graphical databases , We have the opportunity to start with foreign enterprises at the same time . The global map database market has not yet been finalized , Now the layout is just in time , We should firmly grasp this rare opportunity for development .

Graph database is a database with graph as its basic model and data structure in the field of computer science . With excellent expression ability 、 Visualization and solid mathematical foundation , The graph is already in Physics 、 chemical 、 biological 、 Computer science and many other fields have been widely used . Just take the field of computer science as an example , The diagram is used to represent the communication network 、 Data organization 、 Computing flow, data flow, etc , Including artificial intelligence computing framework . Compared with traditional relational database , Graph database is better at dealing with the relationship between data 、 The advantages of easier visual display and more fully meeting the needs of a variety of actual scenarios .

What is graph database ？

Figure database is an online database management system , Used to perform creation 、 Read 、 Update and delete data in the graph data model . Graph databases are often transactional systems .

Unlike other databases , Graph database gives priority to the relationship between data . This means that the application does not need to rely on the outside world to deduce the correlation between data , Or rely on external systems to process data , Such as MapReduce.

By simply abstracting nodes and relationships into connected structures , The graph database makes it possible to build complex models that are closer to the real problem domain .

For graph database technology , There are two important features ：

• This map stores

  Some map databases use this map to store , It is a storage pattern specially designed for the characteristics of graphs , Other databases use relational or object-oriented databases to store graph data . Non local graph storage has great potential performance bottlenecks , Especially when the data size and query complexity increase significantly .

• Graph processing engine

  This map processing is the most efficient means to process map data , Because the data connection is physically saved in the graph database . Other non graph databases are handled in other ways , Cannot optimize for graph data structures CRUD operation .

The advantages of graph database

Graph database is specially designed and built for processing highly connected data . This is the era of big data , Data ownership has a huge scale and a high degree of internal relations , The graph database is used to realize “ Sustainable competitive advantage ” Provides opportunities .

Apply graph database to real problems , It is mainly determined by the following advantages of graph database ：

Performance improvement

Query performance and response speed are the primary concerns of many customers when establishing data platforms . Online trading system , Especially large Web Applications , Must be done in the shortest possible time （ millisecond ） Respond to user requests , To ensure that customers do not leave because of impatience . In a relational world , As application datasets grow in size , The limitations of complex connections are beginning to show , Performance declines rapidly .

The development cycle is significantly accelerated

Figure data model solves the impedance mismatch problem that has plagued the field of software development for decades , This reduces the development cost of switching back and forth between the object model and the table relational model . meanwhile , Figure data model eliminates impedance mismatch between technology and business areas . Customers in various fields can use the shared graph data model to discuss and describe the core business areas , Then incorporate it into the application itself .

Respond quickly to business

Successful applications rarely stay the same . Business conditions 、 Changes in user behavior and technology and operational infrastructure drive new requirements . In the past , Requires organizations to perform detailed and long-term data migration , Including modification mode 、 Transform data and maintain redundant data to support both new and existing functions .

Using graph database development is fully in line with today's flexible test driven development practices , Allow the graph database to evolve in sync with other applications and any changing business requirements . Data teams can... Without compromising existing functionality , Add the new graph structure to the existing graph database , Instead of trying to model the database in detail and perfectly from the beginning .

Enterprise database platform

In key business applications , The database must have strong scalability , And usually need to support transactional . Although the graph database is relatively new 、 Not as good as RDBMS mature , But there are also graph databases that provide all the functions needed by today's large enterprises ：

• ACID transactional

• High availability

• Horizontal scalability of read operations

• Store billions of entities

These characteristics are important factors to promote large enterprises to adopt graph database . Figure databases are not just medium-sized 、 off-line 、 Or department level applications , It can really change the business model of the whole enterprise .

Figure ten application fields of data

“ big data ” It's growing every year , But today's business leaders not only need to manage larger amounts of data , There is also an urgent need to derive insights from existing data . that , How should CIOs and CTOs gain these insights ？ This requires enterprises to abandon the practice of simply collecting data points , Start building relationships between data . The relationship between data points is even more important than the individual points themselves . Graph database can not only effectively store the relationship between data points , And it's very flexible , Suitable for adding new relationship types , And make the data model adapt to the new business requirements . Except for the database itself , Graph technology also covers the emerging field of graph data science , Suitable for predictive analysis and machine learning , The graph data visualization used is suitable for data discovery and exploration for specific purposes . So what fields can graph databases be applied to ？

The following describes in detail the ten application fields of graph database calculation .

Fraud detection

Banks and insurance companies lose billions of dollars every year due to fraud . Traditional fraud detection methods often can not minimize these losses , This is due to performing discrete analysis , Vulnerable to false positives and omissions . Graph technology provides new methods , Through advanced context analysis , Be able to accurately detect fraud gangs and other complex scams . This enables the graph database to prevent advanced fraud in real time .

When it comes to graph based fraud detection , You need to enhance fraud detection through link analysis . therefore , Two things are clear ：

• As business processes accelerate 、 Improved automation , The time interval for detecting fraud is shrinking , The demand for real-time solutions is becoming more and more urgent .

• Traditional technology is not suitable for detecting complex fraud gangs . Graph databases add value by analyzing associated data points .

To develop efficient and easy to manage fraud detection solutions , Graph technique is ideal . In the face of fraud group 、 Colluding with gangs is still a sophisticated criminal , Figure database technology can expose various important fraud patterns in real time .

Real time recommendation engine

The key technology of real-time recommendation is graph database , In linking a large number of buyers and product data , Beyond other database technologies . To make effective real-time recommendations , The data platform must understand the relationship between entities , And the quality and intensity of these associations .

Graph technology can be purchased by users 、 Interactions and comments effectively track these relationships , To provide the most meaningful insight into customer needs and product trends .

Knowledge map

When using traditional keyword based search , The results obtained are highly random 、 The effect is limited 、 Poor quality . Can't really come up with more accurate 、 More useful questions , And get the most relevant 、 The most meaningful information . With the massive growth of data assets , A knowledge map is needed to accommodate the intrinsic relationships of data sets . By using the graph database , It can improve information access , Users and customers can find the products they need most 、 Services or digital assets .

Anti money laundering

Today's anti money laundering (AML) The scheme is becoming more and more complicated , Often involves cumbersome indirect operations , To identify suspicious activities that mislead and cover up fraud . However , Traditional technologies are not designed to span many intermediate steps , Connect those intermediate steps . Inspectors usually spend a lot of time poring over a lot of data , It often takes months , Resulting in a serious accumulation of daily pending transactions .

With the massive growth of data assets , Technical solutions that can adapt to large-scale and diversified associations in data sets have become an urgent need . Graph databases can identify complex data relationships , And has the ability of real-time query , This is a powerful weapon against money laundering and misappropriation of illegal funds . Adopt new data sources and new types , Continuous improvement AML testing , Without rewriting the data model . Built in high availability features ensure your mission critical AML The engine can always access user data . Unlike relational databases , Graphs store interrelated data , Make the detection of fraud activities unaffected by the depth or shape of the data , More convenient .

Master data management

Master data is the lifeblood of an enterprise , Include the following data ：

• user

• Customer

• product

• Account

• partners

• place

• Business unit

Because master data is highly correlated and shared , If MDM The system is not built in a good way , Reduce business flexibility , And affect the whole enterprise . Most traditions MDM The system relies on a relational database , This database is not optimized for traversal relationships or fast response .

With the development of business analysis , These data connections and relationships within the master data set are critical to competitive advantage . The graph database is a model 、 Store and query levels in master data 、 Ideal for metadata and connectivity .

Compared to building a relational solution , Use graph database , Master data is easier to model , resource requirement （ The modeler 、 Architects 、 DBA And developers ） less . Besides , It is not necessary to migrate all master data to one location . Graph relationships can be easily correlated CRM System 、 inventory system 、 Isolated data between accounting and point of sale systems , Provide a unified view of enterprise data .

Supply chain management

In essence , Supply chain management is dynamic , There are many activities , Bottlenecks can occur anywhere . The challenge lies in the lack of real-time data volumes and details generated by traditional databases 、 Accurate information processing capability .

Now , Big data is driving the rapid development of many industries and even all industries , In which the , Supply chain management is a data-driven view 、 A perfect example of predictive analysis and productivity assurance . Because the data will only increase , The challenge of tracking and maintaining control will become greater .

Use graph database , Can achieve 360 Global storage of degree views , So as to effectively manage and utilize supply chain management data association . When the data source changes over time , The graph model is easier to evolve , To achieve flexible performance and easy to scale , No matter how large the dataset is , Can make real-time decisions .

Enhance the network and IT Operation management capability

Same as master data , The graph database is used to gather information from different inventory systems , Provide a single view of the network and its users , It covers everything from the smallest network elements to the applications that use them 、 Service and customers .

The graphical representation of the network makes IT Managers can catalog assets , Realize visualization of asset deployment , And identify the dependencies between the two . The connection structure of the diagram enables the network administrator to perform complex impact analysis , Answer the following questions ：

• In the application 、 service 、 virtual machine 、 The physical machine 、 Data Center 、 Router 、 Switches, optical fibers and other network parts ,  What are the parts that a particular customer depends on ？（ Top down analysis ）

• conversely , If a particular network element fails , Which applications in the network 、 Service and customers will be affected ？（ Bottom up analysis ）

• For the most important customers , Whether the whole network has redundancy ？

The graph database of the network can also enrich the intelligent operation based on event correlation . Whenever the event correlation engine （ Such as complex event handler ） Inferring a complex event from a low-level network event flow , The impact of this event on the graph model will be evaluated , And trigger any necessary compensation or mitigation operation .

Find out 、 Capturing and understanding complex interdependencies is essential to effectively run networks and IT The key to operation , This is crucial to running a business . Whether it's optimizing the network 、 Optimize application infrastructure , Or provide more efficient security related access , These problems involve a complex set of physical devices and human interdependencies , This is quite challenging for management .

The relationship between network and infrastructure elements is rarely linear or purely hierarchical . however , Graph database is created to store these interrelated data , This makes the network and IT It becomes very easy to transform data into actionable insights .

Data lineage

Risk modeling is a multiple combination of requirements . It also requires organizations （ Especially large banks 、 Hedge funds and aggressive investment institutions ） Investing in 、 holding 、 Tracking data connections in a complex network of financial instruments and fine pricing data .

Compared with traditional system , Using a graph database that complies with privacy and risk reporting regulations has the following advantages ：

• Track the evolution of risk factors , Go back to their original authoritative data source

• Will price 、 Hold positions 、 Cash management and other discrete data systems are extended to a unified data set to work with regulators , Visualize and modify the risk model diagram

• Ability to easily modify risk models , To keep up with changing market conditions 、 Organizational change and investment strategies

• Deal with mergers that affect the history and future operations and performance of the trading department 、 Divestiture and reorganization transactions

Identity and access management

Graph database can store complex 、 Closely related access control structure , These structures span billions of participants and resources . The rich and varied structured data schema supports hierarchical and non hierarchical structures , And its extensible attribute pattern allows to capture the rich metadata of each element in the system .

With a query engine that can traverse millions of relationships per second , Large complex graph database access lookup will be performed in milliseconds , It doesn't take a few minutes or hours .

With network and IT Same operation , Figure the database access control solution allows top-down and bottom-up queries ：

• What resources can a particular administrator manage （ Company structure 、 product 、 service 、 Agreement and end user ）？（ From top to bottom ）

• Given a specific resource , Who can modify their access settings ？（ Bottom up ）

•   What resources can end users access ？

Figure the access control and authorization solutions supported by the database are especially suitable for content management 、 Joint authorization services 、 Social networking profiles and software as a service (SaaS) Products and other fields , Compared with previous relational databases, the processing time is a few minutes , The new database dramatically improves performance to a few milliseconds .

List of materials

Most enterprise manufacturers use vendor applications ：CRM System 、 Work management system 、 Accounts payable 、 Accounts receivable 、 sales Point system, etc . Because of this way , The data needs to be stored and modeled as a graph , The interconnection master data stored in the native graph is not a pure linear structure , Nor is it a hierarchy .

Again , burgeoning BOM Trends also bring more layers of data challenges . More stringent compliance requirements . Counterfeit parts flooded the market . Integration is under way among component manufacturers , The life cycle of components is becoming shorter and shorter . meanwhile , Technological innovation has proven to increase availability risk , It will also lead to obsolescence .

The flexible graph data model , The master data can be easily adjusted when the demand changes over time .

At the end

The books listed above are far from exhaustive according to their application fields , But it shows that graph technology has a far-reaching impact 、 Applications with obvious benefits . even so , There are many other examples of graph technology , Including life sciences 、 Social networks 、 game 、 The government 、 Sports and even non-profit organizations .

Today's enterprise managers are facing more and more pressure , When the data set becomes larger and larger and the processing becomes more and more difficult , Still need to provide actionable insights from your own big data . They need technical solutions that can determine the links between data points and draw convincing and reasonable conclusions .

Figure database is the technical solution . Databases allow data professionals at all levels to go beyond a single data point , Tap the potential of data relationships . How to use these relationships , It all depends on the imagination of database users .

thus , Come to the conclusion ： Graph databases are not out of date , But this new trend full of big data insight .

The blue ocean brain map data all-in-one machine has a simplified and highly available cluster architecture . Soft and hard integration , Highly integrated . Open the box . Better than the current centralized storage architecture X3, Higher than centralized storage architecture X5. Professional operation and maintenance platform , In depth monitoring and management all-in-one system . Distributed storage , high reliability , Full architecture redundancy design , Avoid any single point of failure , And cross node data protection , Better serve all walks of life .