June 2019 Summaries

This week's edition of the Neo4j Newsletter covers various topics including modeling provenance data in Neo4j and techniques for de-duplicating ingredients in the BBC Good Food Graph. Featured community member Ray Lukas, an Architect Advisor at CVS Health, shares his passion for connected data and creating customized training materials on Neo4j development. Stefan Bieliauskas shows how to model provenance data using Neo4j, while Lju Lazarevic discusses techniques for de-duplicating ingredients in the BBC Good Food Graph. Additionally, there are updates on GRANDstack starter templates, ObservableHQ notebooks, and a revamped Neo4j Developer guides. The newsletter also highlights an interview with David Meza, Chief Knowledge Architect at NASA, about his work building the lessons learned graph, and features a tweet of the week from Janos Szendi-Varga showcasing a live demo during a Neo4j meetup in Budapest.

Hackolade developed a tool to prepare data for storage in MongoDB and other NoSQL databases, recognizing a niche in the market. Data modeling is not necessary in theory for schemaless NoSQL databases, but it can be tricky in practice. Hackolade's tool helps with design of these databases, originally developed for MongoDB and document databases, now adapted for Neo4j to cater to Fortune 500 customers who use both. Neo4j is the leader in graph databases, its adoption is great, and Hackolade has been supported by Neo4j sales engineers and staff. The future of graph databases holds promise with improved engineering problems solved by having an interface to see connections in data, making it easier for business-facing folks to explore data and reducing friction between sides.

In this session, we explore the application of graph database technology in the automotive industry within manufacturing, discussing various data models and use cases such as supply chain management, logistics, and claims processing. The session also touches on the integration of graph algorithms with artificial intelligence and machine learning technologies to unlock value from manufacturing data.

Fan Li, a Principal Data Scientist at DuPont, discusses how his company uses Neo4j graph technology to develop a competitive intelligence platform. The platform aims to provide fast, quality, and deep insights for scientists, engineers, and business users. DuPont's mission is to deliver sustainable and innovative solutions to solve some of the world's greatest challenges. The company has diversified its business into various industries such as nutrition, health, industrial bio-sciences, safety, protection, and chemical manufacturing. Due to the diverse data sources and end-users, DuPont needed a platform that could ingest large amounts of data from external sources and provide clear connections for users with overarching missions. They initially attempted using SQL but found it challenging due to schema limitations and flexibility issues. Neo4j was introduced as an alternative graph database solution, which offered native graph technology, a vibrant open-source community, and flexible data schema. The benefits of using Neo4j include its natural flow for mapping data sources into entities and relationships, easy analysis by end-users, and flexible data schema. DuPont uses Neo4j to build a graph-based platform that includes publications, patents, government funding, and other business data. The platform provides features such as search engines, alert systems, dashboards, interactive analysis, reporting, and self-service analysis using BI tools. The company aims to expand its knowledge base with internal technology data, enable end-users to incorporate their personal knowledge and insights, develop advanced features like recommendations, and incorporate other technology areas. Ultimately, the goal is to connect scientists, engineers, and external collaborators to deliver innovative solutions.

Emil Eifrem, CEO and Co-Founder of Neo4j, emphasizes the importance of integrating context into artificial intelligence (AI) systems to ensure they are ethical, reliable, and robust, responding to a request for information from the United States National Institute of Standards and Technology (NIST) about developing federal AI standards. Eifrem argues that context, derived from connections naturally managed by graph technology, enhances AI system performance and accountability by providing insight into logic processing pathways. Neo4j, a leader in the graph platform category, has been involved in numerous government and commercial projects, demonstrating the effectiveness of graph-based data handling in various applications, including cybersecurity and equipment tracking. Eifrem stresses the role of public-private collaboration in driving innovation and transparency in AI development, advocating for the inclusion of graph technology in technical standards to incorporate comprehensive contextual information, thus maximizing the ethical and economic benefits of AI solutions.

This week, Robert Schäfer and Grzegorz Leoniec showcased how they use Neo4j and GraphQL to build Human Connection, a free and open-source social network for active citizenship. Meanwhile, Andy Robbins visualized Bloodhound data using PowerBI, while Shawn Roberts analyzed the 990 Forms published by the IRS to visualize the relationships in the data related to the college admissions scandal. Additionally, Janos Szendi-Varga and Miro Marchi started a series of posts on monitoring Neo4j with Prometheus, and Emil Eifrem was interviewed on the Open Source Underdogs podcast. Structr version 3.3 was also released, featuring improved security, data deployment, and query functions, while Dean Wilson shared an entertaining tweet about experimenting with neo4j and python to model teams and engineers.

In a lighthearted tone, Neo4j's Chief Scientist Jim Webber shares his expertise on graph databases. The company celebrates Swedish Midsummer, a summer solstice celebration, as a company-wide holiday and invites employees to participate in festivities. To stay updated with new videos, subscribe to the Neo4j YouTube channel for weekly releases of graph technology content.

In this session, we explore how using connected features improves the accuracy, precision and recall of machine learning models by leveraging graph algorithms that provide more predictive features and aid in feature selection to reduce overfitting. A link prediction example is also presented to demonstrate the measurable improvement of graph-based features in inferring collaboration. To access more content like this, viewers can register for access to the Neo4j Webinar library.

CA Technologies, a global information systems company, was undergoing digital transformation when it began incorporating Neo4j. They aimed to cover their extensive data landscape and manage multifaceted problems with different solutions. The company sought a solution that could encompass the customer footprint and user journey from start to finish, but faced challenges due to dispersed data in various locations. To address these issues, CA Technologies tackled three key elements: incorporating people with diverse skill sets into a cross-functional team, developing an interactive and responsive workflow, and rebooting the culture around sharing information across teams. By leveraging Neo4j, a graph database, they were able to model heterogeneous data sources and relationships, enrich their data, and provide a better understanding of their data. This enabled the engineering and data science team to work more efficiently, access data easily, and share it company-wide. The adoption of Neo4j has accelerated digital transformations at CA Technologies, providing business intelligence to teams and delivering value through data-driven insights.

GKE provides an easy-to-deploy and manage platform for running Neo4j, allowing users to deploy high availability graph databases clusters in as little as 10 minutes. The combination of GKE and Neo4j offers technical flexibility and agility, enabling users to customize the database as needed, run multiple databases alongside each other, and handle increased user demand with autoscaling. Additionally, GKE provides security features such as fine-grained network policies and automatic updates, ensuring secure deployment and management of the infrastructure. The use of Kubernetes also allows for cloud portability, enabling developers to move applications between clouds without vendor lock-in. The integration of Neo4j with GraphQL enables users to build data-intensive full-stack applications with minimal effort. Overall, GKE provides a powerful combination of ease of deployment, technical flexibility, and security, making it an attractive option for organizations looking to solve connected data problems with high performance and flexibility.

This week's Neo4j Online Meetup featured modeling and visualizing borehole data with Neo4j and GraphXR, as well as a new sandbox for Bloom, a graph visualization tool. The meetups also covered exploring shell companies using graph algorithms, redesigning the developer mini site, importing data from Google Cloud Storage, and creating a Women's World Cup graph. Featured community member Sony Green was highlighted for his contributions to the Neo4j community, including Kineviz's work on virtual reality graph visualization and their partnership with Neo4j.

Coshx, a technology consulting firm, uses Neo4j to improve performance in client projects, particularly when dealing with database JOINs. They chose Neo4j due to its native graph model, which allows for more expressive querying and reduces the need for JOINs. The company has seen surprising results from Neo4j, including improved performance and better connected data analysis. If they were to start new projects again, Coshx would plan ahead, invest time in whiteboarding, and reach out to the community sooner for help with potential issues. Looking forward, graph databases like Neo4j are expected to make it easier to tackle difficult engineering problems by providing a visual interface to explore data connections.

Hadoop, once hailed as a revolutionary big data platform, is facing significant challenges with an estimated failure rate of 85% for customer projects and vendors like MapR on the verge of shutdown. However, experts believe that Hadoop will continue to play a key role in enterprise data architectures, albeit with a focus shifting from "big data" to "connected data", where relationships and context are more valuable than sheer volume. As graph technology emerges as a way to connect and contextualize data, buyers will increasingly look for solutions that combine Hadoop or other technologies with graph capabilities to achieve success and value in their data projects.

The presentation discusses how metaphors used to describe the world directly correspond to the way technology, particularly artificial intelligence and machine learning, is built. The U.S. Census is cited as an example of how a simple metaphor led to the development of modern computing. The presenter argues that there's a need for robust metaphors to talk about new technological capabilities and notes that experts in the field are struggling to explain AI and machine learning to the public. The presentation highlights various machine learning technologies, including natural language generation, probabilistic methods for real-time streams, image analysis, and summarization. It also touches on the importance of ethics and data science in building responsible AI systems. The presenter emphasizes that it's essential to think about the impact of technology on the world we live in and encourages practitioners to consider the ethics of their work.

Kubernetes is an open source container orchestration platform that enables users to package applications and their dependencies into containers, making workloads portable. It helps manage complex workloads involving multiple containers by orchestrating them according to policies defined by the user. Google Kubernetes Engine (GKE) simplifies the operational overhead of managing Kubernetes by hosting and managing it for the user, providing scalability and uptime guarantees. GKE also enables users to run a wide variety of applications and services, including persistent storage and databases, with high availability. The Google Cloud Platform Marketplace features numerous applications and components that can be easily deployed on GKE, making application deployment easy. Additionally, Neo4j is a highly scalable, native graph database that stores and manages data relationships as first-class entities, enabling users to query their data and discover connections and relationships among data much faster than traditional databases. By combining GKE with Neo4j, developers can solve connected data problems with high performance and flexibility.

This week's Neo4j community newsletter covers various topics including exploring football data with Neo4j R Driver, combining Cassandra and Neo4j to analyze the Wikipedia Graph Dataset, using Google Cloud Run as a serverless backend for a GRANDstack application, and releases of several Neo4j tools such as Neo4j Bloom, Neo4j Streams, Neo4j GraphQL-Java, and Halin. The newsletter also highlights the launch of the first-ever Neo4j Online Developer Expo and Summit (NODES) which will take place on October 9th, 2019, and features a tweet from Uwe Geercken about his interests in Neo4j and Docker.

The Jim Webber Show features engineering insights and British humour from Neo4j's Chief Scientist, covering topics such as graph database scalability, native graph database differences, and weaknesses of graph operations on non-graph data stores. The show has likely addressed many engineering-level questions related to graph technology. For those interested in learning more about graph databases, the show is available in a growing playlist on the Neo4j YouTube channel, updated weekly with new content.

Neo4j Bloom is a graph data visualization tool that allows users to explore and understand their Neo4j graph in a codeless, natural language interface. It provides a simple and intuitive way for non-technical team members to contribute to graph analytics and development teams, while also offering powerful features for advanced Neo4j users to create custom searches and perspectives. The tool runs in an internet browser, making it easily accessible to project teams, and allows administrators to set up role-based access to schema-like views of the graph called perspectives. With Bloom, users can visualize graph traversals, reveal highly connected nodes, and illustrate relationships between different node types, all while maintaining ease and security for database administrators.

The NODES 2019 event will feature a day-long gathering of Neo4j engineers and community members, with interactive technical talks, live chats, and Q&A sessions. The event will include multiple tracks, keynotes, regular sessions, lightning talks, and panel discussions on various graph-related topics, such as graph-powered AI and machine learning, engineering recommendations engines, and best practices in data modeling and Cypher queries. There will also be community activities, including a Global GraphHack, local in-person viewing parties, and "The Hunger Games" - LIVE 5-Minute Challenges with Prizes. The event is open to developers in the Neo4j community, who are invited to submit talks on technical topics. Registration is now open, and attendees can provide feedback during the registration process to help shape the schedule and content of the event.

Graphs are a versatile and dynamic tool used by NASA to solve complex problems in their mission. The organization had been collecting project data since the late 1950s, but accessing it was challenging due to silos between departments and within individual groups, products, and programs. To break down these silos, NASA achieved significant results using Neo4j, a graph database that enabled engineers to identify trends, prevent disasters, and incorporate lessons learned into new projects. The Lessons Learned Database, part of NASA's knowledge management strategy, has already generated significant value by saving at least a year and $2M in research and development. Going forward, the team plans to provide users with the ability to input lessons directly into the database and run text analysis to identify similar documents. This approach allows users to leverage knowledge from the past to improve decision-making and transmit it to the next generation. Graph databases are versatile and can be used by various government agencies to fulfill their missions, and Neo4j helps customers realize the power of graph technology to solve challenging problems.

This week, the Neo4j community explored event-driven graph analytics using Neo4j and Apache Kafka, with a focus on streaming data between causal cluster instances. They also learned how to load data into a remote Neo4j instance, model Instacart data, and build an Incident Response Knowledge Graph. The featured community member, Vlad Batushkov, completed a 1-month graph challenge, creating various graphs and challenges others to do the same. Additionally, the APOC library was released with new features, including functions for working with dates and procedures for merging nodes and relationships. Various other topics, such as modeling events in Neo4j and loading tweets into Kafka and Neo4j, were also covered.