March 2019 Summaries

The GraphTour Europe event has concluded with a break before the US version begins in Washington D.C. and San Francisco. The GRANDstack team demonstrated Neuler - The Graph Algorithms Playground, created a new starter kit, and had various presentations and coding sessions. Pierre Romera, Chief Technology Officer of the International Consortium of Investigative Journalists (ICIJ), was featured as a community member for his work on data journalism and creating transformative tools to enable journalists. The GRANDstack also expanded its startup program, made it easier to apply, and provided free Neo4j Enterprise Edition and Neo4j Bloom licenses to members. Community members shared their projects, including graphing Brexit analysis and property graphs with Elixir. Jennifer Reif updated the Neo4j Knowledge Base and Developer Pages, while Mark announced that he would be taking a break for the weekend.

Enterprises face challenges in adopting graph technology, which represents a paradigm shift for most organizations. This shift is partly driven by a change in mindset, empowering more people to envision the possibilities of graph technology and drive demand from the business upwards. Dan Woods, who spoke at Neo4j GraphTour, emphasized that systematic adoption of graphs requires a number of steps, including finding initial use cases, creating a center of excellence, expanding it operationally, and making it secure. He also highlighted two key factors in enterprise graph technology: enabling a large number of people to understand the power of thinking in graphs and building the necessary plumbing for graph ETL. Woods noted that graph ETL is the skill that will be the boundary for companies as they adopt graphs, and he believes that Neo4j's Morpheus feature will make this process easier. He also discussed the shift from a project-based approach to adopting graphs to using graphs as a platform for technology, and how Google Kubernetes Enterprise and Neo4j can work together to integrate data from microservices into a unified view. Additionally, Woods sees an overlap between graph technology and data warehouses, where graph technology can provide integration of data from various applications and enable the creation of a unified model that radiates value. Finally, he finds it interesting that Neo4j is trying to be both an operational database and an analytics OLAP database, which has not happened in the relational world.

Scripps Networks Interactive is a leading developer of engaging lifestyle content in the home, food and travel categories for television, the internet and emerging platforms. Their U.S. lifestyle portfolio comprises popular television and internet brands that collectively engage over 190 million consumers each month. Managing metadata across multiple digital asset management systems is critical to their business, affecting broadcast availability, viewing behavior, and syndication of content to non-linear and international broadcast partners. The company uses Neo4j to federate digital asset management systems into a cohesive multimedia library, leveraging graph technology to model complex relationships between multiple domains such as shows, seasons, episodes, formats, and versions. This allows for modeling of collections of short-form videos that share different aspects and enables easy querying of relationship data. The use of Neo4j provides the company with abstraction, simplicity, and query elegance, enabling them to handle increasing complexity in their business operations.

This blog series aims to help developers effectively utilize graph analytics and graph algorithms using a graph database like Neo4j. It explores various community detection algorithms, including the Louvain Modularity algorithm, Triangle Count, and Average Clustering Coefficient algorithm. The Triangle Count algorithm measures the number of triangles passing through each node in the graph, while the Average Clustering Coefficient is used to estimate whether a network exhibits "small-world" behaviors based on tightly knit clusters. These algorithms have been shown to be useful in classifying website content as spam or non-spam, investigating community structure in social graphs, and detecting communities of pages with a common topic. The blog series also provides examples and use cases for these algorithms, including a small dataset example using Neo4j's Cypher query language.

The Neo4j Online Meetup series has featured a presentation by the International Salmon Data Laboratory on using Neo4j to track climate change, and this week's meetup is focused on Game of Thrones. My colleagues Irfan and Mark have created a Graph App called Neuler that utilizes graph algorithms for various use cases, including analyzing the Game of Thrones dataset. They will demonstrate how graph analytics can help understand the show without even watching it, using data made available by Professor Andrew Beveridge. The meetup is part of the Neo4j YouTube channel's weekly updates with tons of graph tech goods, and viewers can catch all videos by subscribing to the channel.

The Neo4j community has had another busy week with various activities such as the launch of Neo4j Labs, a one-stop shop for solution offerings that provide fertile ground for innovation. The team also shared research into link prediction and presented a meetup on this topic, as well as showcased how to implement log shipping with Google's Stackdriver. Additionally, there have been updates on various projects including Zendesk integration, Marvel UI, and Geo-spatial analysis. Dr. Alexander Jarasch was featured as a community member, highlighting his work in combining research data sources to develop effective prevention and treatment measures for diabetes. The week also saw the launch of a revamped Neo4j Administration course, which introduces administrators to common administration tasks for a production Neo4j application using Neo4j Enterprise Edition 3.5.

The German Center for Diabetes Research (DZD) is using Neo4j to connect and analyze various data sources related to diabetes, including genetics, epigenetics, and metabolic pathways. Dr. Alexander Jarasch explains that they chose Neo4j due to its ability to handle heterogeneous data and curate it in a human-readable format. The researchers have been impressed by the ease of building data models and querying the database, which has led to faster discovery of connections between different types of data. As graph technology continues to evolve, Dr. Jarasch believes that it will become increasingly important in research, particularly in fields like biology where everything is inherently connected.

With the help of numerous colleagues and partners, Neo4j Labs offers a wide range of solutions designed to improve developers' workflows. The platform allows for rapid innovation and iteration, providing high-value tooling while actively supporting projects with at least one Neo4j engineer. Initial offerings include the Neo4j Graph Algorithms library, GRANDstack for GraphQL development, APOC Library as a comprehensive toolkit for Neo4j, and other projects that can potentially graduate to the Neo4j Graph Platform or be sunsetted if they don't provide significant community value. To get started, visit the Neo4j Labs site and explore available projects to download and try them out, sharing feedback on the Community Site and Forum.

Neo4j Labs is a one-stop shop of solution offerings that provide fertile ground for innovation, allowing Neo4j engineers to iterate quickly and provide high-value tooling. The platform includes projects such as the Neo4j Graph Algorithms library, GRANDstack, and the APOC Library, which simplify data integration, graph refactoring, and operational functionality. These tools are actively worked on and supported by Neo4j engineers and contribute to the online community. Successful projects can graduate into the Neo4j Graph Platform, while those that haven't proven valuable will be sunset and open sourced. Users can explore the latest innovations on the Neo4j Labs site and provide feedback through the Community Site and Forum.

Tom Sawyer Software is a graph data visualization tool that was founded long before graph databases first hit the market. The company has expanded its services to include solutions for network topology, link analysis, and more, serving customers across various industries such as aerospace, retail, and others. By relying on Neo4j as the central tool for their data integration, Tom Sawyer offers a variety of data views and more detailed graph analysis. The company's CEO Brendan Madden discussed the challenges of working with big data and how they're solving these problems using Neo4j. Tom Sawyer has grown into a full-service data visualization vendor over the past few years, offering solutions that span from network topology to link analysis, and customer data models for binding data to multiple sources. The company's design-preview-deploy tooling architecture allows customers to decide where to deploy an application after development, with options including Java, .NET, web or desktop deployment. Tom Sawyer also uses Cypher and Neo4j Bolt protocols for automatic property graph extraction and rapid application development. The company offers a range of data views, including charting architecture, table-based views, tree views, and VB-style inspector views, as well as advanced drawing and layout functions such as constraint-based automatic drawing and labeling technology.

Global Graph Celebration Day is a global event that takes place on April 15th, marking the legacy of Swiss mathematician Leonhard Euler. The day aims to celebrate his contributions to graph theory and data analysis, which have had a significant impact on our society. Euler's work on traversing graphs with exactly one edge per node led to the concept of "degree of nodes" and the idea that any given graph can be traversed with each edge traversed exactly once if it meets certain conditions. The event is organized by Neo4j, a team that approached the community to help start this movement. The celebration will take place in various locations across 5 continents, including São Paulo, London, San Francisco, and more. Participants are encouraged to host their own events or attend existing ones, with resources available on the Global Graph Celebration Day website.

The Louvain Modularity algorithm is a community detection algorithm used to evaluate social structures in networks, including Twitter, LinkedIn, and YouTube. It measures the quality of an assignment of nodes to communities by comparing their relationship density to a suitably defined random network. The algorithm has been applied to various domains, such as recommending subreddits based on user behavior and extracting topics from online social platforms. However, it also has limitations, including a resolution limit that can make it difficult to detect small communities in large networks, and a degeneracy problem where there are multiple community assignments with similar modularity scores. The algorithm has been successfully applied to real-world data sets, including a graph of users and friends, which was used to demonstrate its capabilities. Next week's focus will be on the Triangle Count and Average Clustering Coefficient algorithm.

The Neo4j community has been actively engaged with various projects and initiatives this week. The introduction to Neo4j online course was revamped and is now available for free, covering the basics of graph technology and Cypher queries. Irfan Karaca showcased a Graph App called Neuler that makes it easy to run graph algorithms over data without writing code, while David Allen discussed running Neo4j on Kubernetes. The community also explored geospatial indexing with Uber H3, as well as deleting multiple properties from nodes. Additionally, Neo4j has joined the GraphQL Foundation as a Founding Member, and Joe Chesak was featured as this week's community member for his efforts in starting a local Neo4j meetup group in Norway. The community also participated in a desktop graph analytics meetup where Neuler was showcased, and Sebastian Mueller presented an introduction to yWorks and its integration with Neo4j.

DXC Technology is leveraging Neo4j, a native graph database, to dynamically track industry trends, client roadmaps, and implementation projects. This enables the company to understand complex relationships and influences between these factors, ultimately informing their data-driven solutions. By using Neo4j, DXC has seen significant improvements in development speed, analytics capabilities, and ability to visualize data, leading to new insights and a more agile approach to their work. Looking back, Stevens realizes that the original approach was too traditional and that disconnecting and connecting nodes has led to greater agility. He predicts that graphs will continue to transform the way organizations approach client conversations and understanding industry trends, unlocking new views and enabling recommendations, knowledge management, and better investment decisions.

There are many facts and figures about tea that could be represented as a graph, including details about countries that commercially produce tea and information about their tea production. For each country, there are many facts and figures that could be included in a graph, such as population and geographic coordinates, and details about its tea production, including total yearly volume of tea produced, number of tea estates, etc. The graph could then break down each country into its tea producing regions, with new nodes to represent additional detail in the geographical hierarchy, such as counties or prefectures within each province, and areas, cities, or towns within each county. Tea itself is also a key part of this graph, with six main types of tea that have different production processes, and details about their recommended brewing parameters. A knowledge graph like this could be used to represent and work with a large amount of knowledge about a particular domain, such as tea, and could power search engines, make recommendations, or provide fast access to complex and highly connected information for cutting-edge technology solutions like chatbots and machine learning.

SocialBee is a software tool developed by Novetta that uses social network analysis to identify trends and patterns in online communications. It pairs metadata with data and communication content to analyze behavior, sentiment, and topics over time. SocialBee can predict hidden relationships, entity resolution, and clustering of communities based on behavior. The tool uses Neo4j to store data and applies non-negative matrix factorization (NMF) topic modeling to extract meaningful insights from unstructured text data. By incorporating structured text into a social network stored in Neo4j, SocialBee can identify influential nodes for specific topics and find disconnected nodes within communities. The tool has been tested on the Enron email corpus with promising results, including uncovering hidden relationships and entity resolution challenges.

The GraphQL Foundation is a neutral group established to provide governance and vendor-neutral stewardship for GraphQL, aiming to grow the ecosystem through educational content, events, and technical advisory committees. The foundation includes founding members such as Neo4j, Apollo, AWS, IBM, Twitter, Hasura, Facebook, Intuit, and Paypal, collaborating with the community to encourage wider adoption. GRANDstack is a framework for building full-stack applications with GraphQL and Neo4j, offering a bundled project with all necessary features, and enables clients to query data as a graph, resulting in better performance and developer productivity wins when used with a graph database like Neo4j. The integrations utilize GraphQL type definitions to drive the database schema and API, generate a single Cypher query from each request, and extend GraphQL with Cypher queries through schema directives, providing improved performance and additional functionality.

Valentine's Day became Graph Visualization Day, with many people sharing their favorite graph visualizations using Neo4j. The posts showcased a wide range of applications, from game states and bridges in the United States to genetic variants, security issues, and even chocolate products. Users also shared their own projects and experiences with Neo4j, including data visualization, machine learning, and IoT applications. The community came together to celebrate graph visualization and share their love for Neo4j, a graph database that enables fast and efficient querying of complex relationships between entities.

This week was a busy one in the Neo4j community, with various online meetups and blog posts covering topics such as blockchain, link prediction, healthcare search, and more. Dr. Lena Wiese, Head of research group Knowledge Engineering at Georg August University Göttingen, was featured as a community member for her work on graph databases and presented a tutorial on data analytics with graph algorithms. Thomas Silkjær and Sony Green showed how to build a graph of the XRP ledger, while I started a series of posts on link prediction, including an overview of the problem, several link prediction measures, and the challenges of building a link prediction machine classifier. Other notable posts included leveraging graphs for healthcare search and explaining what happens when you query a Neo4j cluster using one of the official drivers.

Kineviz, a San Francisco-based company, uses artistic cues to surface complex information in high-dimensional datasets, aiming to make data visualization more intuitive and user-friendly. They specialize in visualizing and presenting complex data, often integrating their product with Neo4j to leverage its graph database capabilities. According to Weidong Yang, CEO of Kineviz, working with data is an art rather than an engineering skill, requiring intuition and artistic understanding. The company's integration with Neo4j has made it easier for them to focus on visualization and interactivity while leaving hard data problems to the graph database. Graph technology is promising but still in its early phase, and there's a need for research to make graphs more accessible and useful in various applications.

GraphXR is a progressive web app that enables intuitive exploration of graph data in virtual reality and traditional 2D computing environments. It offers features such as statistical link analysis, geospatial visualization, timeline filtering, and the incorporation of rich content like portrait images and in-app video. GraphXR can be installed from the Graph Applications tab on the Graph App platform, which is a natural fit for users who need to conduct fast and intuitive workflows. The app runs in a web browser or Electron shell, with user data staying in their database and only login credentials being sent to the server. A free Explorer edition supports Neo4j Community Edition, while the full-fledged Graph App ecosystem aims to democratize graph databases across various industries and applications.

The Label Propagation algorithm is a fast algorithm for finding communities in a graph, detecting these communities using network structure alone as its guide and not requiring a predefined objective function or prior information about the communities. The algorithm works by propagating labels throughout the network and forming communities based on this process of label propagation. It has been used to assign polarity of tweets, estimate potentially dangerous combinations of drugs to co-prescribe to a patient, and infer features of utterances in a dialogue for a machine learning model. The algorithm results in different community structures when run multiple times on the same graph, but can be narrowed down by assigning preliminary labels to some nodes. It has been found to have diverse applications across various domains.

This call for participation is for the Visual Software Analytics research group from Leipzig University, which is participating in the Google Summer of Code program. The group is looking for independent participants to work on various projects related to graph-based software analytics and visualization, with a focus on open-source development using tools like Neo4j, A-Frame, and Unity. Participants will receive a scholarship of $5,000 and have three months to complete their project, with the option to work on hardware such as HTC Vive or Oculus Rift. The projects range from classic software development to designing virtual reality applications, and participants can choose from a variety of tasks and mentors, including those working on jQAssistant and Getaviz.

Bea Hernández shared her talk on football and graphs at SatRDays, discussing how to model football matches in a graph and analyze home advantage and competitiveness. The Kafka Connect Neo4j Sink Plugin was released in the Confluent Marketplace, enabling streaming of events from Apache Kafka into Neo4j. Michael Hunger presented on analyzing StackOverflow data using the Neo4j Graph Algorithms library, including computing tag correlations and running other graph algorithms. Jhonathan de Souza Soares is this week's featured community member, recognized for his engagement in the Neo4j community and his work translating the Definitive Guide to Graph Databases into Portuguese. The GraphTour EU conference continues with stops in Stockholm and Berlin, while Amy Hodler discussed her upcoming book on graph algorithms and AI on the Graphistania podcast.

Phunware uses Neo4j as the engine to power a knowledge graph that connects over a billion nodes, allowing them to deliver great mobile experiences and reach users with relevant content at scale. Their dataset is huge, with over 1.5 billion nodes and 17 billion relationships, and performance was critical for them, with query times in milliseconds or seconds. Neo4j has enabled Phunware to uncover data quality issues and improve their data quality, as well as bring together data from hours or days into seconds. The company's Phunware Knowledge Graph is a key part of this, allowing customers to build upon the dataset and extend it with domain knowledge, enabling them to deliver personalized experiences with mobile device profiles. Overall, Neo4j has been instrumental in helping Phunware achieve their goals, and they are excited about its potential for future growth and development.