January 2022 Summaries

This week we used a subset of a published NFT Trades dataset to model, import, and analyze in Neo4j AuraDB Free. We started with a tweet about research on NFT trades, which led us to the authors' data, a 6.1M trades CSV file. After importing the data into Neo4j AuraDB Free, we created a graph model with nodes for NFTs, traders, and transactions, and relationships between them. We then analyzed the data using queries such as counting the number of trades, finding the top 10% of traders who account for 85% of transactions, and identifying the highest profit made by a trader on a given day. We also used Neo4j Bloom to visualize the data, exploring co-buying behavior, collections of NFTs, and transaction volume and cryptocurrency. The analysis revealed interesting insights into the NFT market, including pure sellers, buyers, and traders who trade frequently with each other. This was just a small taste of what can be done with this data, and we encourage readers to explore further and come up with their own queries and visualizations.

GraphConnect will be back this year in June, with a call for papers open until March 25. The event aims to bring together the global graph community and encourages submissions on various graph topics. Neo4j is looking for experts and has open positions available globally. A 2021 Neo4j Developer Survey is also ongoing, with participants eligible to win goodie bags. Maxime Guery, a.k.a Cobra, is this week's featured community member, known for his positive influence in the Neo4j community. GraphConnect will feature various topics, including AI and data science, applied professionally through his role at Data Nostra. The event will also showcase examples of how to use the Neo4j JavaScript Driver and create a clinical knowledge graph using Spark NLP and Neo4j. Additionally, there's a tweet of the week featuring Dilane Kombou's appreciation for a graph database made with Neo4j.

This tutorial outlines how to build a JSON RESTful API using Neo4j, PHP, and OpenAPI. It starts by setting up Neo4j AuraDB, a managed cloud solution with automatic backups, high availability, and scaling built-in. The author then installs the Slim framework, creates a controller for managing users, and defines routes for CRUD operations. The API uses Cypher queries to interact with the graph database. The tutorial also covers advanced topics such as relationships, idempotency, and querying relationship distance. Throughout the process, the author emphasizes the simplicity and ease of use of this technology stack.

Graph data science is a way of using graph data to gain insights and make predictions. It's about understanding the connections between entities in your data, rather than just looking at individual pieces of information. Graphs can represent complex relationships between data points, making it easier to identify patterns and trends that might be missed with traditional data analysis methods. By representing your data as a graph, you can uncover hidden insights and make predictions about what's likely to happen next. Neo4j is a platform that supports graph data science, offering tools and algorithms for analyzing and visualizing graph data. With Neo4j Graph Data Science, you can use graph algorithms and machine learning models to gain deeper insights into your data and make more accurate predictions.

The organization Jackson Laboratory, a nonprofit biomedical research institution, used Neo4j to manage and analyze large amounts of genomic data. The researchers chose Neo4j due to its ability to scale with their data, efficient graph querying capabilities, and robust REST API for interacting with the graph. They built two prototypes, one using SQL databases and another using Python, but found that Neo4j's bulk import feature significantly reduced the time it took to build the graph. The researchers successfully integrated Neo4j with Google Storage and deployed a Community Edition instance on Compute Engine. Today, their graph contains around 3.6 billion relationships between genes, transcripts, variants, eQTLs, and other data sources. The graph is used by researchers to analyze genomic data, retrieve tables of links between genes, human genes getting chromosomes, and genes linked to variants in another species without an ortholog. The team plans to integrate the graph with geneweaver.org and deploy it externally. Neo4j's scalability, query efficiency, and REST API made it an ideal choice for their use case.

A solution is presented for combining Natural Language Processing (NLP) with knowledge graphs to extract valuable insights from relevant articles automatically. The proposed approach uses Diffbot's APIs to retrieve articles related to cryptocurrencies, translates foreign articles using Google Translate API, and imports them into Neo4j. NLP models are then used to extract entities and relationships from the text, which are stored in Neo4j for graph analysis. Various applications can be developed using this data pipeline, such as monitoring news about companies or competitors, predicting future events like acquisitions, and fueling machine learning models for desired use cases like predicting crypto trends. The code is available on GitHub, and the solution has been applied to a sample dataset of articles related to cryptocurrencies, resulting in valuable insights into entities, relationships, and sentiment analysis.

The Neo4j Data Importer is a client-side browser tool that allows users to load CSV files, create a graph model of labeled nodes and typed relationships, map CSV files to the elements of the graph model, run the import against local and remote servers, and export the model into shareable datasets. The tool is available for non-encrypted server connections hosted on an http URL. It can handle medium complexity data imports with 10 node- and relationship-types and medium size data (100k to 1M elements). Users can also use post-processing to handle some aspects that the data importer doesn't cover, such as filtering out unwanted fields or converting units like quantity to integer. The tool is designed for initial data imports and is not suitable for complex data transformations or recurring import runs. It supports denormalized datasets and allows users to create self-relationships and run incremental modeling. Users can provide feedback on the tool and learn more about it in GraphAcademy courses.

The text discusses the exciting future of Web3 and its connection to property graphs, which will form the basis of Web4j. It highlights various topics such as NFTs, SmartContracts, web development with PHP, sports data science, and Neo4j-related projects like Orai, Embeddings, Named Entity Recognition, and Sports Science, as well as tutorials on building a web app with Neo4j AuraDB and PHP, and using Chainlink to interface with Neo4j from within an Ethereum smart contract. The text also features a featured community member, Cristina Escalente, and invites readers to participate in the 2021 Neo4j Developer Survey.

Emil Eifrem, CEO of Neo4j, shares his experience on what it takes for a company to become category-leading in the cloud infrastructure industry. He emphasizes the importance of a virtuous cycle framework that consists of four stages: awareness, activation, fit/user-centricity, and growth. The awareness stage involves creating community programs and engaging with developers to generate excitement around the product. Activation is about making it easy for users to find success with the product. Fit/user-centricity involves evolving the product to meet the needs of as many users as possible. Growth is about sharing user success, which drives in the next wave of users. Emil also provides four bonus tips for success: being everywhere, offering a free tier that's greater than a free trial, redrawing differentiation between features and usage, and keeping friends close by adjusting open source licenses and creating feature differentiation. Note: I've omitted some details to condense the summary into one paragraph while maintaining its key points.

Neo4j, a graph technology company, has experienced immense growth in 2021, including a new funding round and board appointment. The company's products have been featured in various media outlets, highlighting their ability to excel at spotting data relationships at scale, preventing cyberattacks, providing critical business insights through knowledge graphs, and helping organizations harness big data. Neo4j's CEO, Emil Eifrem, has emphasized the importance of graph technology in addressing complex problems such as climate change and financial corruption, positioning the company for success in 2022.

Doctor.ai is a voice virtual assistant powered by graph database Neo4j and AWS, designed to provide healthcare-related functionalities. It uses natural language understanding engine Lex and Lambda function to query the Neo4j database, mapping large amounts of medical records to a graph and recommending treatments based on similarity with other patients' lab results. The system has potential for improvement, including enhancing voice interaction, expanding treatment recommendation capabilities using knowledge graphs like Hetionet, and exploring open-source alternatives due to AWS reliance. Despite its limitations, Doctor.ai demonstrates that functional medical chatbots are within reach, offering a possible bottom-up adoption pathway for individual patients or doctors.

This edition of Twin4j covers a wide range of topics, including data science and developer tools for Neo4j. Samuel Longley advocates for ETL as the most fun way to load data into Neo4j, while Tomaz Bratanic shares tips on optimizing data fetching from Neo4j with Apache Arrow. David Stevens analyzes social network graphs in NeoDash to identify common groups between people. Meanwhile, Sixing Huang has built a virtual voice assistant called Doctor.ai using Neo4j and AWS, which has potential applications in healthcare. The community is also highlighted through the featured member Rhys Evans, who models complex services as a graph stored in Neo4j and accessed through GraphQL. Additionally, Jennifer Reif explains how to transition from relational to NoSQL to graph databases in her video on Neo4j Live - Code Wars: Database Decisions for Application Development.

The conversation between Jim Webber and Dr. Edgar Osuna revolves around the use of graph technology in real-time financial crimes detection by TODO1, a company that services banks. Dr. Osuna explains how Neo4j is used to identify bad actors, with the system performing well even when faced with high volumes of transactions and complex patterns. The commercial side of Neo4j has been instrumental in supporting TODO1's work, with the team from CMS managing their cluster and environments, providing peace of mind for the company. Dr. Osuna also discusses how emerging economies have a competitive advantage over developed economies due to lower costs and higher levels of education, allowing them to participate more actively in the technology industry. He highlights the importance of understanding these differences when building businesses that operate globally.

The Neo4j blog has highlighted several key topics from 2021, including the Pandora Papers dataset and its use of graph databases to analyze financial secrets. The blog also covers various tools for visualizing Neo4j graph databases, such as Graph Data Visualization, and provides a one-stop solution for importing CSV files into Neo4j. Additionally, a tutorial on constructing a biomedical knowledge graph using NLP techniques is shared, while the company announces that Certified Neo4j Professional t-shirts are available with the purchase of the certification test. The Neo4j Certified Professional exam tests skills in Cypher, basic graph data modeling, and multi-database features, among other areas.

Google Cloud Vertex AI and Neo4j are used together to build graph-based machine learning models. Graphs describe the relationships between data points, making them suitable for handling complex business problems. Google Cloud and Neo4j offer scalable tools for leveraging graph data. The PaySim dataset is used as an example to identify fraud in credit card transactions. Data scientists can use basic graph algorithms or generate graph embeddings to encode complex graph data into values that tabular machine learning algorithms can use. The embeddings are then combined with Vertex AI to train a supervised model. Future work includes exploring improved datasets, boosting and embedding methods, and manually engineering new graph features.

As Neo4j's chief scientist, a report on knowledge graphs was written to explore the concept of digital twins and their potential for businesses. A digital twin is a virtual representation of a physical object or process that serves as its real-time counterpart, but it can also be applied to complex systems of software and people. Digital twins have been used in various industries such as power generation, transportation, and pandemic response, often in combination with knowledge graphs. However, implementing digital twins poses challenges due to the complexity of modern systems and the limitations of traditional data technologies like relational databases. Knowledge graphs can help overcome these challenges by providing a more efficient way to process and query complex interconnected domains. A case study from Lending Club demonstrates the potential of digital twins on top of knowledge graphs, enabling real-time insights into system performance and behavior. The report also highlights vertical opportunities for building digital twins in various domains such as asset tracking, maintenance, and financial flows.

The Orai application is a project management tool that uses the Neo4j graph database and Django web framework to build sprint cycles for development teams. It monitors tasks dependent on each other, prioritizes them based on complexity, and allows users to control the length of sprints, essential tasks, and the number of tasks per sprint. The system takes in four user inputs: max complexity value per sprint, minimum number of sprints desired, a set of prioritized stories, and an optional maximum dependency tree length. It provides example roadmaps based on inter-story relationships, helping teams plan and schedule their projects more efficiently. The team behind Logic13 developed Orai as part of Florida Atlantic University's Senior Design Program, learning new programming skills and collaborating through GitHub.

Neo4j is an embedded graph database that offers added value and functionality to applications, enhancing their performance, security, and reliability. It provides a broad set of drivers for popular programming languages, can be deployed in various topologies, and is exceptionally efficient at processing connected data. Neo4j uses Cypher, a powerful open graph query language, which enables developers to turn complex SQL queries into simple ones. The database has a thriving global community, providing access to a large ecosystem of extensions and examples, as well as a global talent pool. It boasts an open library of user-defined procedures and functions, and its flexible data model allows organizations to add data easily without schema redesign or downtime. Neo4j is a proven technology used in more enterprise deployments than any other graph database, offering performance at scale, security, fault tolerance, and high availability. Its capabilities enable organizations to respond to events within an application in real-time, rather than after the fact. The database protects data and information, providing fine-grained access control for robust management of users, roles, and data access. Neo4j is a recognized leader in graph data platforms, offering a complete graph data platform that includes visualization tools, development environment, data science library, native machine learning capabilities, language APIs, and data integrations. The company has been downloaded over 100 million times, making it the most adopted graph database by developers and enterprises, enabling developer velocity to build modern, intelligent applications at scale.

In this post, we explore how to build practical and scalable recommendation systems using Neo4j and its Graph Data Science library. We start with a fundamental example of news recommendation on a dataset containing 17.5 million click events and around 750K users, leveraging Neo4j's graph data model to quickly predict similar news based on user preferences. The post defines the basics of Recommender Systems, explains how they work, and provides an overview of the Microsoft MIND dataset used in this example. We also cover a technique called Collaborative Filtering (CF) which is applied using Cypher query language and scaling with the Graph Data Science library, leveraging node embeddings and an ML technique called K-Nearest Neighbor (KNN). The post concludes by discussing next steps and providing resources for further exploration of graph-based recommenders.

Neo4j has released several exciting updates in 2021, including Neo4j 4.4, which provides significant performance improvements and reduced memory requirements for large datasets. The company also introduced a new GraphAcademy platform, offering free online training with updated courses and a redesigned look to make learning graph database technology more accessible. Additionally, the community has been actively engaging in discussions on various topics, including data modeling, graph embeddings, information extraction pipelines, and machine learning applications using Neo4j. The company is currently conducting its 2021 Neo4j Developer Survey, inviting participants to share their experiences and opinions on the latest developments in the graph database space.

Neo4j, a leading graph database platform, has had a remarkable year in 2021, despite the challenges posed by COVID and climate disasters. The company raised $325M in investment, introducing new features and updates in its latest version, Neo4j 4.4. To support its growth, Neo4j introduced GraphAcademy, a free online training platform, and expanded its Certified Professional program with new t-shirts and designs. In the developer community, Neo4j released tools for graph visualization, including 15 favorites categorized by functionality. The company also collaborated with data scientists to explore applications of graph databases in machine learning and artificial intelligence. Additionally, Neo4j integrated seamlessly with external libraries like DGL for deep learning tasks and provided a seamless integration with Python for data scientists. The platform continues to support the development of knowledge graphs, enabling businesses to better understand their data in context. With its commitment to innovation and community engagement, Neo4j is poised for continued growth and success.

Neo4j has released several exciting updates in 2021, including Neo4j 4.4, which offers improved performance and data processing capabilities, while also reducing memory requirements for large datasets. The company has also introduced a new version of its GraphAcademy online training platform, featuring updated courses and a redesigned look and feel to make learning graph database technology more accessible. Additionally, Neo4j has released Neo4j Bloom 2.0, which enhances the power of graph visualization, and AuraDB Free, a free Graph Database as a Service that provides users with their own dedicated graph database for free. The company is also currently conducting its 2021 Neo4j Developer Survey to gather feedback from developers and offer participants a chance to win some goodie bags.