June 2022 Summaries

Memgraph, a startup that has surpassed the typical lifespan of two to five years, is celebrating its sixth anniversary with a special video by founders Dominik and Marko, who share their journey while visiting San Francisco. The video offers insights into how the founders met and started the company, along with some lighter moments, such as a playful mention of who is the better dancer. Viewers are encouraged to engage further by joining Memgraph's Discord for questions or exploring career opportunities on their website.

Zeljko Riha shares an enthusiastic announcement about a review of Memgraph by data scientist and researcher Ashleigh Faith, who presents topics related to knowledge graphs and data science on her YouTube channel. The review features Memgraph's DevRel Engineers, Katarina Šupe and Ivan Despot, and is part of Ashleigh's Knowledge Graph Technology Showcase series. The blog encourages viewers to subscribe to Ashleigh's channel and explore Memgraph's capabilities through various platforms, including the Memgraph Playground, an open-source Memgraph Platform, and a trial of Memgraph Cloud. Additionally, readers are invited to engage with the Memgraph Community to discuss projects and explore data streams.

Art Blocks is a digital platform that uses Ethereum-based NFTs to integrate the creation and ownership of algorithmically generated art. To explore the dynamic nature of these transactions, Memgraph has developed real-time analytics tools using streaming data, showcased on their Awesome Data Stream website. By using Memgraph Lab and Memgraph Cloud, users can query and visualize this data to gain insights into NFT transactions, identify popular projects, accounts, and sales, and detect communities within the dataset through graph algorithms like PageRank and community detection offered by the open-source MAGE repository. The blog post demonstrates how to efficiently update datasets in real-time, ensuring the community detection algorithm performs optimally by implementing dynamic updates and triggers. Users are encouraged to explore the Art Blocks dataset and other datasets on Awesome Data Stream, with the possibility of engaging with the Memgraph community on their Discord server.

The blog post explores the creation of a real-time book recommendation system using the Amazon Books dataset, utilizing Memgraph's graph-based tools and algorithms for dynamic data processing. The process involves setting up Memgraph Cloud or a local instance, connecting to data streams, and using Memgraph Lab for querying and visualization. The system employs the PageRank algorithm to identify popular books and Python query modules to calculate average ratings and generate personalized recommendations by excluding books already rated by users. Additionally, the dynamic PageRank algorithm is introduced for efficient real-time updates in growing datasets. The blog provides a comprehensive guide for building recommendation engines with a focus on scalability and personalization, also encouraging further exploration of available datasets for similar projects.

Memgraph is revolutionizing data analysis by providing developers with tools to extract real-time insights from data streams using graph algorithms, which were traditionally accessible only to big tech companies. By introducing Memgraph Cloud and Memgraph Lab, the company has facilitated a seamless, user-friendly environment where developers can leverage graph algorithms like PageRank and Community Detection on streaming data to gain insights for various applications, such as drug discovery and fraud detection. Memgraph's support for streaming technologies like RedPanda, Kafka, and Pulsar, as well as static data, ensures flexibility and accessibility for developers looking to harness the power of graph databases. The new cloud-based solution eliminates the need for complex installations, offering an intuitive graphical user interface through Memgraph Lab, where developers can easily connect to data streams, execute algorithms, and visualize results. This development aims to accelerate application development by enabling faster, more efficient data analysis and decision-making processes.

The blog post explores how to analyze a streaming dataset of movie ratings using custom query modules in Memgraph, a graph database platform. It highlights the use of the MovieLens dataset to dynamically identify the best and worst-rated movies, as well as the most popular ones by applying the PageRank algorithm. The process involves setting up Memgraph using Memgraph Cloud or a local instance, writing custom Python query modules to update movie ratings in real time, and using database triggers to ensure data consistency. The blog also demonstrates how to retrieve and display movies based on their ratings and popularity, offering insights into both highly acclaimed and less favored films. Additionally, it provides guidance on leveraging Memgraph's MAGE library for efficient data processing and encourages further exploration through tutorials and community engagement on Discord.

GitHub, often seen as both a software development platform and a social network, offers a dynamic dataset that can be analyzed in real-time using streaming data and graph algorithms like PageRank and community detection through tools such as Memgraph. The blog explores the use of these algorithms to identify influential users and detect communities within GitHub's social structure, highlighting the challenges of running computationally expensive algorithms on constantly changing datasets. To address these challenges, dynamic versions of these algorithms, which update only affected nodes, are utilized via the MAGE graph algorithm library. The process involves setting up Memgraph to ingest GitHub data streams, running queries to analyze the data, and creating triggers to automate updates as new data arrives. This approach allows for intuitive and efficient real-time data analysis, making it possible to derive strategic insights from the network's social dynamics.

Graph Data Zagreb is hosting its fifth meetup on July 6, 2022, featuring Michaël Ughetto, a graph data scientist from AstraZeneca, who will discuss the use of the Biological Insights Knowledge Graph (BIKG) in accelerating drug discovery. Ughetto will explain how AstraZeneca integrates data sources into BIKG and how this tool aids scientists in developing therapeutics. The event will take place at WESPA Spaces in Zagreb and offers an opportunity for networking over drinks and burgers, with nearly 200 graph and data enthusiasts expected to attend. Ughetto, who has a background in particle physics and machine learning, is focused on applying graph technologies to drug discovery challenges, and the meetup encourages participation from both existing and new members of the group, with additional engagement opportunities available on Discord.

On May 8, 2022, the fourth edition of the Graph Data Zagreb meetup took place, attracting over 50 participants interested in cryptocurrency and blockchain technology. Boris Agatić, a data scientist and tech entrepreneur, delivered a captivating presentation demonstrating how graph databases can be used to analyze real-time transactions on the Bitcoin network. The event fostered engaging discussions among attendees, who enjoyed burgers and beer afterward. The announcement for the next meetup, scheduled for July 6, 2022, features Michaël Ughetto from AstraZeneca, who will discuss the use of graph technologies in drug discovery. Participants are encouraged to join the Graph Data Zagreb group to stay updated on future events and engage with the community on platforms like Facebook and Discord.

In-memory databases (IMDBs) utilize RAM for data storage, offering significant speed advantages over traditional disk-based databases, making them ideal for applications requiring rapid response times such as gaming, streaming, and real-time bidding. These databases store data in a non-relational and compressed format, using data tiering to differentiate between frequently accessed data (hot storage) and less critical information (cold storage), which enhances data retrieval speed and efficiency. IMDBs are particularly beneficial in data science applications due to their ability to handle big data with reduced IT overhead and costs, while allowing for fast queries and real-time data processing, which aids in machine learning model training. Despite their advantages, they are susceptible to data loss in events like power outages, although techniques such as transaction logging and data snapshots mitigate this risk. Tools like Memgraph, Aerospike, Hazelcast, Redis, and SAP HANA offer various features to leverage in-memory databases for data science by facilitating real-time data processing and advanced analytics, supporting diverse data structures, and ensuring security through encryption technologies.