January 2016 Summaries

Aseem Kishore, the founder of FiftyThree, discusses how his company uses Neo4j in its Paper app for capturing and sharing ideas. He highlights the benefits of using Neo4j, including its ability to handle complex interconnected data models and support high write throughput. Aseem shares some interesting results he's seen while using Neo4j, such as creating a tree-like structure to convey remixes between ideas, without denormalizing the data or precomputing the list. He also reflects on what he would do differently if he were to start again with Neo4j, including breaking out disparate types of data into separate nodes to improve performance and scalability. Overall, Aseem is excited about the latest developments in Neo4j and its potential for supporting complex graph-based applications.

Zephyr Health is integrating diverse healthcare data using MongoDB and Neo4j databases. The platform derives real-time analytics that help patients connect with different therapies and pharmaceutical companies connect with different healthcare providers. The architecture of Zephyr's platform was built from scratch in Java and Groovy, utilizing ontology-driven development to integrate unknown and unseen data in a more consistent way. A polyglot persistence database is used, combining MongoDB and Neo4j to store data points and profiles as JSON documents in MongoDB and nodes and edges in Neo4j. The system uses fuzzy match logic to combine entities with similar information and stores the documents and profiles as JSON documents in MongoDB. Zephyr's platform has proven flexible, extensible, and adaptive, addressing challenges related to data modeling and organization, such as indexing and data size limits. The graph database technology used by Neo4j enables exploration of relationships that were previously unknown, making it a vital tool for the healthcare industry.

The Neo4j community has shared various articles, blog posts, videos, slides, presentations, libraries, code repositories, podcasts, and audio content in December. These resources cover a range of topics such as graph databases, data modeling, indexing, migration, recommendation engines, Twitter influencer analysis, and more. The community also features user-generated content like GraphGists and code repositories showcasing various use cases and applications of Neo4j. Additionally, the community is promoting its free "Neo4j in Production" class to help users improve their graph database skills.

The new release of Neo4j OGM (Object Graph Mapping) introduces several exciting features, including revamped connection mechanisms with drivers, performance improvements, and bug fixes. The drivers mechanism provides an abstraction over different types of connections to Neo4j, making testing easier and allowing for the use of embedded instances as well as remote instances. Performance improvements include a refactored Cypher statement generator that supports multi-statement queries, resulting in significant speed boosts when persisting large object graphs. Additionally, the new release introduces support for mapping custom query results to domain entities, which will make it easier to deal with the results of custom queries and improve performance. The update also includes several bug fixes, including retrying HTTP requests and fixing connection leaks.

Graph database technology is able to detect patterns that arise around e-commerce fraud scenarios and put an end to them in real-time, before a fraudster can inflict significant damage. E-commerce fraud often involves large numbers of users with transactions originating from the same IP address, shipments using the same credit card, or multiple credit cards using the same address. The pattern inside the graph, discovered by walking relationships between disparate pieces of information, serves as strong indicating signals of an e-commerce fraud event. Graph databases are designed to carry out pattern discovery in real-time across these datasets and can uncover schemes before they inflict significant damage, with triggers including login, placing an order, or registering a new credit card.

The World Economic Forum's Scott David is the head of user experience and design, who led the organization's transition to a graph database, specifically Neo4j, for search and real-time recommendations and transformation maps. The switch was driven by the need for relevance engineering and personalization in their digital products. With Neo4j, they were able to create a unified metastore of knowledge, join experts and projects together, and provide a new way of thinking about concepts and patterns. The choice of Neo4j was influenced by its simplicity, ease of infrastructure setup, Cypher query language, and modeling time in its data model. The project has brought about significant value, including change in thinking, improved business conversations, and exciting results such as building out a global issues taxonomy. Scott David emphasizes the importance of communication of the value of graph technology and watching the growth of the graph space, which is converging with machine learning to create powerful opportunities for enterprises.

The ConceptNet dataset is a semantic network built from nodes representing words or short phrases of natural language, with relationships between them. It aims to enable systems to take human words as input and use them to better search for information, answer questions, and understand user goals. The concept can be applied in various domains such as retail, social startups, and content platforms, where tag-based recommendations and categorizations are used. By translating the ConceptNet5 API into a graph database, it takes away the agony of tag-based recommendations and provides insights on how to categorize products effectively. The dataset is incredibly rich, providing various "senses" in which someone might mean a word and offers a wide variety of relationship types to choose from.

The discussion revolves around microservices architecture, its benefits, and how to implement it effectively. It highlights the challenges faced by traditional organizations when moving code into production quickly, such as water-Scrum-fall, and how microservices help overcome these issues. The talk introduces the concept of bounded contexts, which allows for formal network boundaries between services, enabling polyglot persistence and making it easier to choose the right database technology for each service. It also discusses the importance of hypermedia and REST APIs in describing application state and navigating through applications. Additionally, the presentation touches on challenges such as service registration discovery, configuration, and gateways, and how tools like Zookeeper, ETCD, and Netflix's Eureka help address these concerns. Finally, it showcases a demo on Neo4j deployment using Lattice and Spring Cloud, demonstrating polyglot persistence for microservices.

The demand for Neo4j skills is high, making it a great time to showcase expertise. The first official Neo4j Certification Exam has been launched, offering a way to prove experience with the graph database. The exam consists of 80 questions that must be answered within an hour and require a score of 85% or higher to pass. The certification covers various areas such as graph database basics, Cypher syntax, application development, and graph data modeling problems. Passing the exam earns a certificate, which can be shared with others to demonstrate skills, and also includes adding the certification to LinkedIn profiles for added credibility.

Insurance fraud is a sophisticated and effective crime that can be challenging to detect, but graph databases can help combat collusive fraud by analyzing relationships between individuals involved in fake accidents. Typically, insurance fraud rings involve multiple roles such as doctors, lawyers, body shops, and participants who work together to stage false accidents and claim soft tissue injuries, known as "whiplash for cash." These rings often use complex schemes involving multiple people playing different roles, making it difficult for traditional relational databases to detect them. However, graph databases can be used to identify fraud rings by querying intricate connected networks, allowing insurance companies to flag suspected fraud in real-time and prevent costly claims.

We're excited to announce the schedule for this year's Graph Processing devroom at FOSDEM 2016, featuring presentations on graph database internals, new frameworks, and applications. Petra Selmer and I will introduce the openCypher project, while speakers will discuss graph processing using Apache Flink, large-scale analytics, and data mining. The event also includes a get-together for the graph database community to exchange ideas over a Belgian beer.

The presentation by David Makogon from Microsoft Azure and Patrick Chanezon from Docker discussed how to automate cloud deployments of Neo4j using Docker and Azure. The authors compared public cloud services to movie comparisons, such as Amazon's service being like "Pacific Rim," where extraterrestrial monsters invade the enterprise market. They also compared VMware to the movie "300," which is courageous but ultimately fails. Microsoft's cloud service was likened to the movie "Field of Dreams," where building a public cloud service will attract users. The authors then delved into Docker, explaining how it provides a portable approach to cloud technology and enables DevOps without lock-in. They also discussed how Docker has changed the industry by providing a new stack with operating systems like CoreOS, Red Hat Project Atomic, and Ubuntu Core. In this new stack, there are tools for orchestration such as Docker Swarm, Apache Mesos, and Google's Kubernetes. The authors also introduced the concept of Open Container Initiative (OCI), which standardizes the runtime and bundle format for containers. They explained how to use Docker to build, run, and ship applications, including using Docker Compose for microservices-based applications. Additionally, they discussed Azure Resource Manager, a tool that allows users to create templates for infrastructure, storage accounts, networks, and virtual machines in a single operation. The authors demonstrated how to use Azure Resource Manager to deploy Neo4j on top of Docker, showcasing the automation and efficiency of this approach.

US banks are losing tens of billions of dollars annually to first-party fraud, which is estimated to be as much as one-quarter or more of total consumer credit chargeoffs in the United States. The magnitude of these losses is due to two factors: first-party fraud is difficult to detect because fraudsters behave similarly to legitimate customers until they "bust out" and disappear; secondly, the exponential nature of the relationship between the number of participants in a fraud ring and the overall dollar value controlled by the operation makes it particularly susceptible to graph-based methods of fraud detection. Graph databases can help detect and mitigate three types of fraud: first-party bank fraud, insurance fraud, and e-commerce fraud. Traditional methods of fraud detection are geared towards discrete data and not connections, leading to false positives and undesired side effects in customer satisfaction and lost revenue opportunity. Entity link analysis using a graph database can uncover collusions of the type described above with a high probability of accuracy, especially when running checks during key stages in the customer and account lifecycle such as at account creation, investigation, credit balance threshold hits, or check bounces.

The 5-Minute Interview: Karl Urich, DataFoxtrot` Karl Urich's first Neo4j project was four or five months ago, where he used Wikipedia data to solve a problem involving both spatial and graph data. He chose Neo4j because it is free, easy to use, and has good support and integration with other technologies. Karl found the power of using a purpose-built tool for analyzing graph and relationship-based data to be particularly useful, allowing him to achieve success and experiment with the technology. If he could start over, he might consider formal training, but notes that Neo4j was so easy to use that he didn't see any need to do anything differently. After attending a conference talk on graphs, Karl expressed his enthusiasm for Neo4j and encouraged others to share their own experiences in future interviews.

The World Economic Forum is a meta-organization that utilizes graph databases to analyze global issues, facilitate collaboration among leaders and experts, and provide strategic focus for its members. The forum's annual meeting in Davos brings together 2,500 global leaders to discuss pressing global issues, and the organization has developed various tools to help leaders navigate complex information and identify key trends and patterns. These include transformation maps, which display the relationships between different global issues, as well as graph search and recommendations capabilities that enable users to find relevant content and connect with experts in their network. The forum's use of graph databases is still evolving, but it has the potential to revolutionize the way leaders understand and address complex global challenges.

Graph databases are increasingly being used to detect and mitigate sophisticated scams, including fraud rings, by leveraging connections between data points to uncover patterns that traditional methods may overlook. These databases can analyze connected data in real-time, making it easier to identify and stop fraudulent activity before it causes significant damage. By reframing the problem as a graph, companies can gain new insights from their existing data and significantly improve their fraud detection capabilities. Graph databases offer a unique ability to uncover various important fraud patterns, including collusions that were previously hidden, and are becoming an essential tool for efficient and manageable fraud detection solutions.