December 2018 Summaries

Neo4j, a leading graph database company, had a remarkable year in 2018 with several significant announcements and achievements. The company introduced Neo4j Bloom, a graph data visualization tool, and released Neo4j 3.4 with its Multi-Clustering feature, which allows for the creation of multiple domain-specific database clusters. Additionally, Matt Casters, creator of Kettle, joined the Neo4j team as Chief Solutions Architect, and the company launched its Graphs4Good program to connect and showcase graph-powered projects working towards a better world. Neo4j also closed an $80 million series E funding round, making it the largest single investment in the graph technology category. The company continued to grow and expand its community efforts, including the launch of new forums and recognition programs for local graph leaders. Overall, 2018 was marked by significant milestones and a growing momentum for Neo4j as it takes on the world of data in 2019.

The Neo4j YouTube channel has released a diverse collection of videos on various topics related to graph technology in 2018, including deep learning on graphs, transforming enterprise AI at scale, graph visualization with Neo4j, and exploring the World Cup graph. These videos feature experts from Neo4j and other organizations, such as Andrew Jefferson, Hilary Mason, Ryan Boyd, Will Lyon, Dr. Jim Webber, Jesús Barrasa, Amy Hodler, and Jennifer Reif, who share their insights on topics like cloud and AI, digital transformation strategies, data structures and algorithms, ontologies in Neo4j, and graph algorithms. The videos are designed to be accessible to a wide range of audiences, from beginners to experienced developers, and offer practical advice for implementing new technologies and architectures built on Neo4j and enriched by AI algorithms.

The Graphistania podcast has featured excellent episodes in 2018, showcasing the dynamic world of graph technology. Rik's interviews with Karin Wolok from Neo4j, Irene Iriarte Carretero from Guosto, Michael Simons from Neo4j Engineering, and others highlight various applications and developments in the field. The podcast also features a unique interview with JEP, the Graph Database, and its CEO Emil Eifrem, offering insights into the future of graph technology. These top 10 episodes represent a mix of exciting projects, innovative uses, and expert discussions, making them must-listen for anyone interested in graph databases and related technologies.

Neo4j, a graph database company, is grateful for the engagement with its blog in 2018 and plans to continue providing updates on the latest developments in the field. The blog covers various topics related to graph databases, including beginners' guides, microservices management, civic empowerment through data-driven storytelling, and more. Neo4j's funding has enabled it to expand its offerings, setting up the company for delivering more value to its customers.

The author has shared a holiday gift to the graph theory community, featuring a video on Euler's formula explained by 3Blue1Brown. The video showcases how planar graph duality can be applied to prove Euler's Characteristic Formula, highlighting the math that underlies graph technology. This video provides a unique perspective on graph theory and its applications in solving problems in new ways. The author encourages viewers to subscribe to the Neo4j YouTube channel for regular updates on graph tech content.

This week marks the longest night of the year, and the Neo4j community is celebrating with a well-deserved break. The featured community member this week is Raik Lochau, who recently left his corporate job to start a fashion startup and has been actively contributing to the Neo4j community on StackOverflow. A new online introduction class for Neo4j has also been launched, providing a comprehensive guide to getting started with the graph database. Additionally, several articles and publications have been released, including an article on using Neo4j 3.5's full-text search with GRANDstack's `@cypher` directive, and a blog post by Thomas Silkjær exploring relationships on the ledger using Neo4j. The community is also looking forward to seeing more answers from Raik Lochau in the future, as well as articles and talks on Neo4j-related topics.

KeneteK, a knowledge management company, partnered with LARUS to deliver an enhanced offering leveraging Neo4j graph databases. KeneteK initially used traditional relational databases but found limitations in handling complex business processes and data flow. The introduction of Neo4j enabled them to represent business processes as nodes, connect them to logic/verb nodes, and relate them to other data nodes. This allowed for efficient analysis and reduced project risks by 90%. With Neo4j, KeneteK can now visualize data flows, find cross-silo calls between applications, and connect technical components with people from different backgrounds and business cultures. The partnership aims to bring local and global companies up-to-date and help them achieve digital transformation.

This blog series explores graph analytics and graph algorithms for developing intelligent solutions using a graph database. The Single Source Shortest Path (SSSP) algorithm calculates the shortest path from a node to all other nodes in the graph, which is faster than the Shortest Path algorithm but used for similar problems. SSSP is essential in logical routing such as telephone call routing and is implemented by Neo4j's delta-stepping algorithm, which outperforms Dijkstra's algorithm in sequential and parallel settings. The algorithm assumes positive weights and does not support negative weights. It has applications in Open Shortest Path First, a routing protocol for IP networks, where it helps detect changes in topology. A sample graph is used to demonstrate the SSSP algorithm, showing the cost of going from node A to each other node, including itself at 0. The algorithm is part of a series of pathfinding algorithms, with the next topic being All Pairs Shortest Path.

This week in Neo4j highlights the most interesting developments from the community over the last seven days. Featured community member Joe Depeau, Senior Presales Consultant at Neo4j, launched the Crime Investigation Sandbox, which showcases Neo4j Bloom and POLE investigations. The Applied Graph Algorithms training was previewed, with Will Lyon giving a sneak peek of the course on the online meetup. Michael Hunger wrote a blog post demonstrating how to use Neo4j's graph algorithms to correlate tags in StackOverflow. Jennifer Reif continued her series on building a full stack application with Spring and Neo4j, this time showing how to hydrate the model using APOC library. Pedro Mendonça built a follow system for a social network in Neo4j, while Michael released the first milestone of neo4j-graphql-java - a GraphQL to Cypher transpiler for the JVM. The podcast featured an interview with Will Lyon, and the community is looking forward to next week's developments.

The evolution of graph databases and graph-based systems has been a significant area of growth in the technology sector, with Neo4j remaining at the forefront. The development of a common graph query language, GQL, is underway to standardize this emerging industry standard. Graph technology's maturity and simplicity, particularly the property graph data model, have contributed to its widespread adoption across various industries. Initially developed for content management, graph databases have proven versatile and can be applied to dynamic systems with large amounts of data. Intelligent applications leveraging static and dynamic information are becoming increasingly common, making graphs an ideal representation of real-world structures. The integration of graph technology in artificial intelligence (AI) and machine learning (ML) is also evident, as most models in these disciplines rely on graph-based representations.

The new CFO of Neo4j, Mike Asher, brings a wealth of experience from his previous roles at Chevron, Cisco Systems, and Octane Software. He has worked with high-growth tech startups and has a strong background in finance. Initially skeptical about joining Neo4j due to the company's focus on graph technology, he was convinced by the team's passion, the strength of their product, and the unique blend of Silicon Valley innovation and Swedish employee-centric culture. Asher is optimistic about the future of graph technology in the enterprise, believing it will become increasingly important for companies to manage connected data. In his new role, he aims to operationalize Neo4j's activities, align the distributed team, and drive growth through efficient decision-making and metrics-driven management.

Denise Persson, the Chief Marketing Officer at Snowflake Computing, is excited to join the Neo4j Board of Directors. She brings her extensive experience in technology marketing, having spent over a decade with Genesys and later joining Snowflake two-and-a-half years ago. Denise has a unique perspective on the journey of Silicon Valley startups, where she notes that bravery and vision are essential to succeed. As a board member, she aims to provide fresh advice from her operational experience and contribute to Neo4j's growth and success. Denise is passionate about graph technology and its potential to solve real-world problems, particularly in the connected data space. She believes that diversity and inclusion are crucial for a company's success and wants to help Neo4j attract and retain diverse tech talent. With her new role, Denise's objective is to support Neo4j's mission to make sense of data and accelerate its position in the graph technology world.

Graph technology, invented by Leonhard Euler, is a fundamental field in mathematics and computer science that has been powering innovative tech for over 280 years. It is based on the concept of graphs, which are data structures composed of nodes and relationships between them. Graph databases are online systems capable of storing and managing graph data models, with two key properties: native graph storage and graph processing engines. There are three types of graph databases: property graphs, hypergraphs, and triple stores, each with its strengths and weaknesses. Property graphs are the most widely used type, suitable for rapid traversal of data connections, while hypergraphs excel at capturing meta-intent and RDF triple stores are proficient in offline analytics. Understanding these differences is crucial to choosing the right graph technology for specific applications.

The Shortest Path algorithm is used for finding directions between physical locations, such as driving directions, and also to find degrees of separations between people in social networks. It calculates the shortest weighted path between a pair of nodes, using Dijkstra's algorithm as one of its most well-known variants. The algorithm has a long history dating back to the 19th century and was first implemented by Edsger Dijkstra in 1956. Shortest Path is commonly used in web mapping tools like Google Maps for providing driving directions, and also in social networks like LinkedIn to find mutual connections between people. A real-world example using Neo4j demonstrates how to run the Shortest Path algorithm on a sample graph.

The video series "Intro to APOC" by Michael Hunger is being showcased again, this time with the Editor-in-Chief highlighting its relevance. This week's featured talk is a fun GraphTour 2018 San Francisco presentation on graph algorithms and "Game of Thrones," offering an entertaining way to learn about these concepts. The video explains how graph algorithms are used to identify relationships between nodes, weigh their importance, and provide concrete insights into the world of graph technology. The Neo4j YouTube channel offers a weekly update with various graph tech-related videos, including this one, which can be accessed by subscribing to the channel.

This week the Neo4j community shared various interesting things, including talks and presentations about different topics such as Neo4j Spatial, importing data from APIs like Marvel, and migrating data from Cassandra to Neo4j. The featured community member for this week is Parichay Das, a Technology Evangelist with extensive experience in Data Science and AI. The community also explored the use of Neo4j Bloom for graph visualization, created a full stack application with Spring and Neo4j, and showcased new features in Neo4j Drivers, including Transaction Config and custom server address resolver. Additionally, they discussed Cypher Query Log Analyzer, a tool to help users understand query log files, and shared resources for learning more about Neo4j Spatial and other topics. The next week's events are also announced, along with a tweet of the week from Michael Porter encouraging users to play around with data migration tools.

Thomson Reuters has been collecting data for over 150 years and identified two major challenges faced by its financial analysis clusters: data silos and the lack of a tool to easily uncover important data connections. To address these issues, they developed an intelligent recommendation engine using a knowledge graph, a graph ETL, and Neo4j, which enables financial analysts to more effectively analyze data and make real-time decisions. The solution was built by listening to customer needs, identifying siloed data as a challenge for CTOs and data overload as a problem for financial analysts, and developing a user-centric analytics tool that links, stitches, and joins data together in the knowledge graph stack. This stack includes a knowledge graph with 50,000 employees' metadata linked to core entities, a graph ETL to extract and transform data, and Neo4j as the analytics platform to deliver insights. The recommendation engine uses Neo4j's BOLT protocol with Cypher to load and match data from the data fusion layer, run shortest path calculations, and provide answers to financial analysts' questions without giving them information overload.

Venture capital has become a crucial source of financing for startups, providing necessary financial backing in exchange for a share of ownership. However, with the high risk involved, investors must be selective and thorough in identifying startups with high-growth potential. To overcome this challenge, startups can leverage data visualization tools like GraphXR from Kineviz to analyze investor networks and identify key characteristics such as connections within broader networks, industry investments, and historical patterns of success. By using these visualizations, startups can strategically come up with a more reasonable list of potential partners and gain a better understanding of their network data in new and unique ways.

The yWorks team participated in a Hackathon where participants had to create interactive visualizations of their Neo4j data using various graph-related buzzwords and technologies. The event attracted many developers from diverse industries, and the yWorks team helped teams with visualization tasks without writing any code, providing consultancy services instead. The "Spatial Graph App" team led by Craig Taverner created a custom app that integrated with Neo4j Desktop, using yFiles for visualization and ReactJS for creating the application. Despite initial challenges with the React toolchain, they were able to present their data set to the jury, albeit not winning the prize for best visualization. However, the team won a special prize for their outstanding perseverance and agility. In the end, the "BrowserBuddy" team won the price for Best Graph Visualization, creating an impressive diagram around website domains and http sub-paths using yFiles.

The Neo4j Graph Algorithms Library is a tool designed to help users leverage graph analytics for faster innovation and intelligent solution development. The library offers various graph traversal algorithms, including breadth-first search (BFS) and depth-first search (DFS), which have been tuned for efficiency and streamlined for management and debugging. The library exposes these algorithms as user-defined procedures called as part of Cypher statements running on top of Neo4j. Graph projection is a handy feature that places a logical subgraph into the algorithm when the original graph has the wrong shape or granularity, enabling fast caching for the topology of the graph. Two types of projections are available: label and relationship-type projection, and Cypher projection, which uses node-statement and relationship-statement to project subsets of the graph. The library also offers huge graph projection for large graphs with limitations. Various algorithm types cater to different use cases, including transactions and operational decisions, as well as broader views of patterns and structures across all data and relationships.

This week saw the release of Neo4j 3.5, which includes features such as full text search, indexed backed order by queries, and off-heap transaction state. The first version of the Kafka Connector was also released, allowing data to be communicated between Neo4j's event listener and a Kafka topic. Jennifer Reif started a series of blog posts on building a full stack application with Spring and Neo4j based on the Marvel dataset. Tom Michiels, this week's featured community member, is a data-architect/scientist at Vectr.Consulting and has contributed to the APOC library. The online meetup covered iterative modeling of corporate resources in a rapidly growing company using El Dorado, a tool built by Ward Cunningham to collect metadata from various operations. Jasper Blues started writing a series of posts for young entrepreneurs on topics such as finding traffic routes in good conditions and restoring Neo4j backups.