August 2017 Summaries

Amazon's success in the retail industry is largely due to its ability to leverage technology, innovation, and data-driven decision making. Retailers with declining sales face a major problem of organizational silos that have developed over time from historical workflows. These silos lead to slow decision making as information flows through different teams at different rates, resulting in bottlenecks. To overcome this, retailers need to rethink how they interact with their data and use it to inform decisions. Neo4j graph technology can provide a competitive advantage by harnessing the power of connected data for real-time analytics. This enables product traceability, data-driven decision making, and real-time product recommender systems, ultimately improving customer experience and loyalty.

This week in Neo4j has seen Karin Wolok, Community Manager at Neo4j, being featured for her work organizing and running the Developer Zone at GraphConnect Europe 2017. She has been busy organizing various events, including a GraphQL meetup and meetups with Salesforce and DigitalOcean. The community has also shared their favorite tools and libraries that make their lives easier when working with Neo4j. In terms of learning resources, Ernest Chestnut Jr has written an excellent introduction to graph databases with Neo4j on the Sharp Notions blog, while Laura Kahn has shared her experiences learning Neo4j over the summer. Additionally, Joseph Guhlin released the Omics Database Generator (ODG) which uses Neo4j's graph database as its backend. The community is also looking forward to upcoming events and meetups, including a Graph of Thrones - Neo4j + Game of Thrones event in São Paulo and presentations by Dom Davis and Mark Bogaty at the Neo4j London meetup.

Graphcool has released its GraphQL command-line tools, which can be easily installed using `npm install -g graphql-cli`. The tools utilize the `graphql-config` format for storing endpoints and configuration information in YAML or JSON. For testing purposes, a quick demo is provided to showcase the available commands, including `graphql init`, `graphql ping`, and others. To use the tools effectively, it's essential to set up authentication headers correctly, as an error was encountered when attempting to run `graphql ping` without this crucial information. The author of the blog post delved into the code to find a solution, ultimately covering plugin options, schema fetching, and altering current schemas in a follow-up article on Medium.

The presentation by John Swain of Right Relevance and Microsoft Azure analyzes Twitter conversations around Brexit and the 2016 US Presidential election using graph algorithms. The analysis reveals that social media influencers, rather than a small group of special individuals, play a crucial role in spreading information. The "Law of Quite a Few" debunks the common Internet trope of "the Law of the Few," which suggests that only a few influential people can spread information through networks. Instead, ordinary influencers with certain skills are responsible for spreading information. The analysis also highlights the importance of community detection algorithms to identify groups of people sharing similar values and interests. By using graph algorithms, such as PageRank and betweenness centrality, the team was able to analyze the Twitter data and identify key influencers and topics in the conversation. The presentation demonstrates how Neo4j's user-defined APOC procedures can be used to process large datasets and run graph algorithms natively, improving performance and scalability.

Neo4j is being used by financial services firms to create a 360-degree view of customer experience, enabling them to understand who their customers are, what they're interested in, and how the company can serve them better. This requires harnessing the power of connected data and using graph technology to integrate online channels with existing business operations. By doing so, firms can gain a competitive advantage without disrupting day-to-day business, as seen in a case study of a Fortune 500 credit card provider that integrated Neo4j-powered Master Data Management (MDM) into its organization, resulting in a more complete view of its merchant, partner, and customer ecosystems.

This week in Neo4j highlights the community's achievements and upcoming events. Michelle Sanver, a software developer at Liip, is featured as this week's member of the Neo4j community. She has been actively contributing to the community since 2013, presenting on PHP and Neo4j, and creating the recipe site omnomhub. The community also discusses various topics such as thinking in graphs, importing data from OWL format into Neo4j, and analyzing event data. Additionally, there are several upcoming events scheduled for next week, including a Game of Thrones special, presentations by Dave Fauth, Will Lyon, Michael Hunger, and Pat Patterson, among others. The community is also active on social media, with tweets discussing various topics related to graph databases and Neo4j.

As season 7 of Game of Thrones is progressing, interest in its data is increasing, with several data sources available, including an API created by Joakim Skoog that provides cleaned and scraped data from various online sources. To create a Neo4j graph database based on this data, Michael Hunger has developed a project that utilizes the data to analyze characters and houses in the Game of Thrones universe, using tools such as GraphQL and Cypher. The resulting database is documented in detail and can be used for further analysis or exploration, offering insights into the complex relationships between characters and houses in Westeros.

The author compiled a list of graph theory and algorithms resources, including books, videos, and online materials, to assist with their ongoing project. The list is divided into two sections: "Basics" for those getting started with graph concepts, and "Getting Serious" for those with more mathematics background or ready to dive in deep. The author has reviewed many of these resources but still has a growing list, acknowledging that there may be other notable resources they have missed.

The ICIJ has announced a fellowship program, the Neo4j Connected Data Fellow, to help investigative journalists make sense of complex data and find stories inside networks. The program aims to strengthen reporting by leveraging graph database technology, which was instrumental in ICIJ's Pulitzer Prize-winning Panama Papers investigation. Manuel Villa, a seasoned journalist with expertise in finance, journalism, and data, will join the ICIJ's Data & Research Unit for six months to work on world-changing projects using Neo4j, a graph database technology. The program is designed to help journalists learn Neo4j and become successful with graph database technology, and ICIJ is looking forward to incorporating Manuel's expertise to continue exposing wrongdoing in cross-border investigations.

Neo4j community member Eve Freeman is featured as the community member of the week. She is a long-time Neo4j expert and has authored content for various Neo4j training courses. Eve also created the AnormCypher Scala driver and is a Neo4j Ambassador. This week's online meetup focused on versioning graphs using an entity-state model, with a presentation by Marco Falcier and Alberto D'Este. The project Neo4j Versioner Core provides procedures to help manage the entity state model. Various articles were published on topics such as fake news detection, threat hunting, Symfony integration, and graph visualization. Additionally, there are discussions on StackOverflow about finding triplets, dynamic relationship types, and property arrays in Neo4j. Upcoming events include a presentation on Blockchain as a Graph at the Graph Nerds of Boulder meetup and Real-time Recommender Systems Made Easy with Neo4j at the London Neo4j meetup.

The presentation covers various tips to improve the performance of Cypher queries in Neo4j. Using official up-to-date Bolt drivers is essential, and object mappers should be avoided due to their overhead and limited suitability for batch imports. Query parameters can significantly speed up query execution by allowing caching and reuse of query plans. Reusing identifiers within queries prevents the query plan from being cached, so incremental identifiers should be used instead. Long Cypher queries should be split into smaller, more optimized queries for ease of profiling and debugging. Checking schema indexes is crucial to avoid unnecessary database hits and improve performance. Batch writes using Neo4j's `UNWIND` feature can significantly improve write performance. However, query replanning can slow down queries and introduce higher rates of garbage collection, especially in production scenarios, where disabling replanning through configuration may be necessary.

This post discusses how to leverage graph technology for a solution that aims to gauge the overall controversiality of an article by analyzing its connections with other articles. The proposed framework uses Neo4j's native graph database architecture, which excels at leveraging connected data through index-free adjacency, pattern-matching searches, and graph traversals. By storing entities and topics mentioned by articles as nodes in the graph, users can intuitively search for similar articles using Cypher, a graph query language. The post also outlines the construction of a news graph, which focuses on important relationships between authors, sources, and articles to measure controversiality and identify communities that consistently agree with one another.

This week in Neo4j highlights the community's efforts and new developments in the graph database world. Samuel Yang, a prolific StackOverflow replier, is featured as this week's community member. The release notes for various Neo4j versions highlight improvements such as IPv6 support for clusters and enhanced import capabilities. The first public release of the Neo4j graph algorithms library has also been made available, offering users more tools to analyze complex data structures. Additionally, new releases include a Neo4j desktop version and various online meetups, webinars, and articles discussing topics like Game of Thrones analysis and security engineering with Neo4j.

Glidewell Laboratories, a vertically integrated company with over 4,500 employees, was struggling with an obscured manufacturing process that caused efficiency challenges in developing personalized dental-related products. To address this issue, they adopted a polyglot persistence approach using Neo4j for the live workflow and MongoDB for versioning and metadata. The benefits of the Neo4j workflow engine include its generic, modular, and reusable nature, which allows it to accommodate complex workflows with conditional processes and failure steps. The company was able to improve efficiency, consistency, and visibility in their manufacturing process, making it more transparent and data-driven. They also implemented a unique user view system, allowing managers to track product progress and technicians to submit missing tasks, leading to improved productivity and accuracy.

The Neo4j graph algorithms library is a new public release that provides a set of user-defined procedures for graph analysis on top of Neo4j. The library includes algorithms such as centrality, partitioning, path-finding, and graph projection, which can be used to gain insights from connected data. The algorithms are designed to be efficient in terms of resource utilization and are available in two variants: one that writes the results back to the graph and another that streams the results for further processing. The library is well-documented and includes examples, and it has been tested on various datasets with promising results. The authors invite users to try out the library, provide feedback, and contribute to its development.

The Pew Research Center found that two-thirds of US adults believe fake news has left them confused about basic facts, with a quarter admitting to sharing fake news themselves. Social media is often blamed for the spread of fake news, and an analysis by BuzzFeed News showed that top-performing fake news articles on Facebook resulted in more shares and engagement than major news sources like the New York Times. Experts argue that the problem requires human judgment and cannot be solved solely through technology. Instead, machine learning algorithms can be used to flag potentially misleading news, which is then reviewed by humans. Graph database technology has been shown to be effective in recognizing and leveraging connections in large amounts of data, and could be used to build a "news graph" that provides additional insight into the bias and credibility of individual articles and their sources.

The IT organizations in the financial sector face significant cybersecurity challenges due to the complexity of their data centers, which consist of thousands of servers and multiple physical and virtual networks. To effectively mitigate these risks, they need a technology that can map and visualize this data effortlessly. Graph technology, such as Neo4j, provides a solution by allowing IT professionals to track cyberattacks across their entire infrastructure in an intuitive way, identify single points of failure, and prioritize exposed vulnerabilities. This enables them to strengthen cybersecurity and protect mission-critical systems while keeping business running as usual. The use of Neo4j at organizations like the MITRE Corporation provides insight into the impact of cyber activities and helps analysts to gain visibility into operations for global situational awareness.