December 2022 Summaries

The graph technology landscape is expected to continue growing rapidly in 2023, with major milestones including the adoption of graph databases as a first choice for businesses solving complex data problems. Graph data science and machine learning will also play a significant role, enabling more accurate predictive models and unlocking new insights from graph-structured data. The growth of graph technology will be driven by its ability to handle complex domain models and provide valuable benefits such as improved scalability, efficiency, and resilience. As the industry continues to evolve, Neo4j's Chief Scientist Jim Webber expects to see significant advancements in graph computing, with a focus on making it more accessible and user-friendly for businesses of all sizes. With the rise of Asia-Pacific markets, particularly India and Indonesia, Neo4j is well-positioned to capitalize on this growth and provide its expertise to help businesses unlock the power of graphs.

The Neo4j JSON to C# tool is a useful resource for generating classes for nodes and relationships in graph databases, saving development time. The company also showcased its CloudScreen platform at the NODES conference, which uses machine learning to analyze biochemical knowledge graphs. Developers building modern applications with GraphQL are invited to provide feedback on a new service offering from Neo4j through a short survey. The NODES conference featured various sessions, including a presentation by Fabio Montagna on horizon scanning over a graph of medical research papers and a post by Adrien Sales on modeling an IT system as a social network using Neo4j. Additionally, the company highlighted its Endogamy tool, which uses graph theory to discover inbreeding in family trees, and shared a tweet from Nathan Smith about his experience at the Kansas City Graph Databases Meetup.

Neo4j Ninja at University of Naples Federico II has explored how to combine traditional approaches to artificial intelligence with modern technologies, leveraging Neo4j and the Unreal Engine to create a framework for natural tools and applications with social interactive agents (FANTASIA). This article is part 1 of a two-part series, focusing on how Neo4j supports cross-disciplinary research between technology and humanities researchers. The author discusses the challenges faced by traditional approaches to AI, such as requiring large amounts of data and being difficult to interpret, but also highlights the benefits of combining these approaches with modern technologies. The article describes the development of a new methodology called "Linguistically Oriented Resources and Insights as Expressive Graphs" (LORIEN), which represents linguistic knowledge as graphs connecting multiple resources. This allows for analysis of dialogue dynamics and extraction of general principles that can be applied to various domains, including movie recommendation. The author also discusses the importance of contextual information in understanding dialogue management strategies and introduces a graph analysis tool called HITS, implemented in Neo4j Graph Data Science library, to analyze network characteristics of genre nodes. The article concludes by describing the technological framework developed on the basis of LORIEN structures, including FANTASIA, an Unreal Engine plugin, and Artificial Neural and Graphical Models for Argumentation Research (ANGMAR).

AWS has partnered with Neo4j to help customers use graph databases for various applications, including fraud detection and financial services. The partnership allows AWS to go to market together and provide workshops based on different use cases. According to Roya Niki Hugosson, Strategic Account Manager at AWS, the most surprising results from using Neo4j have been the speed and cost savings that customers can achieve with graph databases. In the future, Roya expects to see more applications of graph technology in areas such as the Internet of Things, IoT, pharma, and developing new medicine and vaccines.

Neo4j is a graph database service that enables developers to build modern applications using GraphQL, and the company is seeking feedback from users through a short survey to inform product design. The survey aims to gather opinions on what features and functionalities users would like to see in Neo4j's new GraphQL service offering. Additionally, Neo4j has made several updates to its tools and services, including an updated Data Explorer with enhanced graph lens capabilities for data visualization, a new Elasticsearch plugin, and other initiatives such as the Graph Academy course on building Neo4j applications with TypeScript. The company also shares news about upcoming events, talks, and resources related to graph technology, including a recent presentation by Owen Brierley on developing a Neo4j plugin for Unreal Engine.

Rijutha Sivaprakas, a Technology Consultant at Ampion, a Wipro company, discussed how her team is using Neo4j to optimize DevOps and migrate to the cloud in the Australian banking sector. The goal was to improve performance and reduce manual intervention in data-driven insights. They built a solution on an existing application but struggled with handling complex, hierarchical, and many-to-many relationships. GraphDB was initially used but ultimately replaced by Neo4j due to its better performance and interoperability requirements. Rijutha found Neo4j's Python driver helpful for integrating with various systems and data sources. Automation of the data ingestion pipeline led to significant productivity gains, reducing manual effort and improving end-user experience. Future use cases include identity and access management in cloud migrations and building decision-making systems for suspicious activities.

The article presents a method for automating the deployment of Neo4j Java extensions to self-managed Google Cloud environments. The current manual process is cumbersome and prone to security breaches, as developers need to share their SSH keys with team members. To address these challenges, an automated pipeline is designed using Google Cloud Platform components such as Cloud Storage, Cloud Functions, and Cloud Run. The pipeline consists of a Cloud Function that monitors the storage bucket for finalized objects and triggers a Cloud Function relay, which then deploys the Neo4j extension to the server hosting the database instance via SSH. The deployment process involves downloading the jar artifact from the storage bucket, deploying it directly via SSH, and restarting the Neo44j DB service. The pipeline can be tested locally using scripts provided in the article. Improvements are suggested for the image spec and deployment, as well as potential modifications to support clustered environments and metadata about deployed artifacts.

The author created a tool to analyze Python project licenses using Neo4j, leveraging PyPi's JSON pages for scraping data. They started with the top 5,000 packages and added an extra 315 due to missing dependencies. The data was stored in a graph database, excluding licenses as separate nodes to avoid finding packages based on their license. To access the data, they used Fast-API and py2neo, requiring five licenses in total. The tool provides a REST API for querying licenses of start packages and all their dependencies with any number of packages in between. Analysis revealed that the most popular license is MIT (42%), followed by Apache (26%), with BSD at 16%. The author concludes that while thinking about licenses might be unnecessary, there is now a tool to help developers manage this.

The Australian Government's Department of Infrastructure is leveraging modern data analytics to improve its transportation policy decisions by harnessing data from various sources, including GPS pings from IoT devices and a national map of roads. By combining these datasets in Neo4j, the department can gain insights into the Australian transportation system, such as identifying areas with high safety risks and optimizing journey paths for drivers. The use of graph technology has enabled the department to analyze vast amounts of data, including 350 million GPS pings per month, and has opened up new opportunities for analytical modeling and policy development.

Graphville is an educational platform for learning Neo4j Database and Cypher query language. The project was in a passive development phase for three years, but the author, Vlad, decided to share his story and summarize the most important aspects of the development process. Graphville aims to help engineers enter the field of graph technologies by providing a fun and engaging learning experience. The platform is based on three key principles: teaching the right things, teaching in a fun way, and setting challenging problems. It features a unique storytelling approach, with each lesson being an episode in a small story with a plot and heroes. The author credits his team, including his sister Alena, frontend engineer, and illustrator Irina, for their contributions to the project. Graphville officially launched on September 1, 2022, and the first season is available, with new episodes being released regularly.

Neo4j is introducing new features in its upcoming update, including a "Show me a graph" preset in Search and a new optional auto-sync feature in Perspectives. The company also released tutorials and examples on how to map out external attack surfaces using Neo4j Bloom and visualize software and infrastructure relationships with Akamai Endpoints. Additionally, Neo4j is showcasing its capabilities through various projects such as the LeakCanary Project, which now supports Neo4j, and a Docker compose example by Christophe Willemsen. These updates demonstrate Neo4j's ongoing commitment to providing tools for developers to build and analyze complex data structures, including graph databases.

Neo4j is a graph database that enables the exploration of connected data, providing insights trapped in rows and columns. With over 70 Data Science Algorithms, Neo4j simplifies machine learning In-Graph, translating connected data into predictive signals through Graph embeddings. This integration with Azure Machine Learning (Azure ML) accelerates and manages the machine learning product lifecycle, enhancing a supervised learning model trained with Azure ML. By leveraging Neo4j's Graph Feature Engineering capabilities, Data Scientists can create in-memory graph projections and node embeddings using Fast RP, exporting them to Azure ML for an AutoML job. The integration enables the creation of an AutoML job to predict policy claims based on enriched data, providing insights into the model and top features by importance. This post introduces Neo4j's integration with Azure ML, showcasing its potential in graph problems across various industries.

Simone Novali, System Architect at Zurich Italia, discussed his experience with Neo4j graph databases in a presentation. He explained how he and his team use Neo4j to manage applications, servers, and costs within their organization. Simone chose Neo4j for its ability to help them achieve their goal of understanding the cost and benefit of their architecture, as well as its strategic partnership status. To get started with Neo4j, Simone advises using people who can define a correct graph, as building an accurate graph is challenging. He also emphasizes the importance of curiosity when transitioning from relational databases to graph databases. Looking ahead, Simone believes that graph databases will become increasingly important for handling large amounts of data generated by various systems and applications, particularly in industries like insurance and IoT. With Neo4j's development of new functionalities and features, especially for distributed systems, Simone sees a promising future for graph databases in digital transformation.

Neanex, a Belgian firm, aims to create a single source of truth for big asset projects by creating digital twins of massive physical objects. They use Neo4j, a highly scalable and flexible graph database, to handle vast amounts of data and manage relationships between them. This solution helps asset stakeholders get a better understanding of what's going on in these complex projects, which are often siloed away with different teams having their own data, tools, and processes. By leveraging Neo4j, Neanex has gained recognition from the European Commission as a Key Innovator and is well-positioned to tackle the scalability challenge of massive construction projects. The company's digital twin solution enables smart, sustainable decisions by connecting people to each other with data, making it easier for teams to make sense of their native data state, whether high-tech or low-tech.

ChatGPT is a conversational AI model that can be applied to various areas such as learning graph databases. It offers several features, including generating information, answering documentation searches, and creating Cypher queries. However, users should exercise caution when relying on its responses, as they may not always be accurate. The model has limitations, such as requiring specific prompts for clear answers and being unable to generate new knowledge. Nevertheless, it can connect disparate parts of existing information, making it a useful tool for understanding patterns and trends in data. Additionally, ChatGPT provides a unique opportunity for creative applications, like generating poems or code examples. To fully utilize its potential, users need to understand its capabilities and limitations, as well as the importance of validation and source attribution.

The presentation discusses Expedia's journey in leveraging graph databases across their brands, from managing multiple database technologies to creating a unified self-service infrastructure for developers. The company created a fully managed self-serve database platform called Cerebro, which allows developers to create high-quality products without managing databases. Cerebro supports single-click provisioning, scalable design, and automated maintenance. By using the Strangler Pattern, Expedia was able to transform legacy applications to use Neo4j/MongoDB architecture, improving security, governance, and productivity. The platform's design uses a root key architecture, allowing for lean storage in MongoDB and efficient querying of Neo4j. This unified approach enables Expedia to unify their platform across multiple brands, providing a single data store for reservations, users, and inventories.

The text discusses various topics related to graph databases, machine learning, and data science, including connection problems in large datasets, the use of Neo4j in web development, and the application of graph embeddings for predictive modeling. It also highlights different use cases such as recommending chemicals-cell interactions, predicting suspicious Bitcoin transactions, and using graph recommendation algorithms for various applications. Additionally, it mentions a tweet from Adrien SALES (@rastadidi) discussing knowledge graphs and Neo4j, as well as presentations and articles on related topics such as Neo4j AuraDB and the power of graph databases.

Neo4j's Data Knowledge Graph was built by Harry Corfield, Senior Data Architect at Telstra, to address the challenge of finding and connecting data across various silos. The project used Neo4j Graph Database to develop a knowledge graph that could bring together different data sources and pieces of information, enabling users to find things more easily. Harry learned Neo4j quickly by taking the company's free Graph Academy courses and certifications, which allowed him to build the tool from scratch in just one week. He also found it useful to import CSV data into Neo4j, create constraints and indexes, display nodes and relationships, and utilize resources like Neo4j documentation, community forums, and Stack Overflow. Harry recommends starting small with a one-week sprint project, showcasing features to gather feedback, and incorporating user requests into the knowledge graph.