COVID-19 Visualization - What Graph Databases Can Do

 Graph visualization intuitively and intelligently shows the structure and association between data

See what you couldn't see in a table or graph before

"A picture is worth a thousand words"

——The editorial department of "New Programmer"

With the advent of the era of big data, traditional relational databases have become increasingly difficult to meet a large number of frequently changing needs due to their limitations in data modeling and storage. Relational databases, despite the word "relationship" in their name, are not very good at handling queries and analysis of complex relationships. In addition, relational databases also lack the ability to scale horizontally over multiple servers. Based on this, a class of non-relational databases, collectively referred to as "NoSQL" storage, came into being, and was soon widely researched and applied.

NoSQL (Not Only SQL, non-relational database) is a wide range and diverse types of data persistence solutions. They do not follow the relational database model and do not use SQL as the query language. Its data storage does not require a fixed table schema, and often avoids the use of SQL JOIN operations, and generally has the characteristics of horizontal scalability.

2021 database panorama, the new database era of "New Programmer"

Comparison of graph databases and relational databases

Graph database refers to a class of databases that represent, store and query data in graphs. The "graph" here has nothing to do with pictures, graphs, charts, etc., but is based on the concept of "graph theory" in the field of mathematics, which is usually used to describe a certain relationship between certain things. For example in our daily life:

Social networks are graphs. The participants of each social network are nodes, and our interactions in the social network, such as "add friends" and "like", are the edges connecting the nodes.

Urban traffic is a graph. Every intersection, house number, bus stop, etc. is a node, and a street or bus line is an edge, connecting the places that can be reached.

Knowledge is also a graph. Each name, concept, person, event, etc. is a node, and the generic relationship, classification relationship, causality, etc. are edges, connecting the nodes to form a huge, rich and evolving knowledge graph.

"Graphs are everywhere", and because of this, traditional relational databases are not good at dealing with relationships, and can be well solved by graph databases, which are born to solve this problem.

Knowledge Graph about the new coronavirus

A group of members from the Neo4j graph database community have integrated multiple heterogeneous biomedical and environmental datasets to build a knowledge graph about the new coronavirus

https://github.com/covid-19-net/covid-19-community

to help researchers analyze interactions between host, pathogen, environment and virus.

Knowledge Graph about the new coronavirus

  • Part of the visualization results of the knowledge map, the leftmost part of the figure is the sub-map of the geographic location of the virus outbreak, including countries, regions, and cities;

  • The middle green part is the epidemiology submap, including information about virus strains, pathogens, and host organisms, with cases and strains associated with where they were reported and found, respectively;

  • The purple part on the right is the biology subgraph, representing organisms, genomes, chromosomes, variants, etc.

 The future of graph databases shines

In finance, graphs and graph analytics help institutions more efficiently spot abnormal connected transactions to win the war on money laundering. In the power and telecommunications industries, graph databases help manage complex and huge equipment and line networks, and analyze the root cause and estimate impact of failures in a timely manner.

In the fields of manufacturing, scientific research, and medicine, graph databases are widely used to store and query knowledge graphs, and have become an important support for big data management, data analysis and value mining, and even artificial intelligence technology.

Database Technology Development Trends (as of June 2021) Source: DB-Engines

In the foreseeable future, the combined application of graph database and artificial intelligence technology will bring more innovations and leaps. Graph databases can help improve AI capabilities in at least the following four areas.

The first is the knowledge graph, which provides domain-relevant knowledge/context for decision support and helps ensure that the answer is appropriate for that particular situation.

Second, graphs provide higher processing efficiency, so using graphs to optimize models and speed up the learning process can effectively enhance the efficiency of machine learning.

Third, feature extraction analysis based on data relationships can identify the most predictive elements of the data. Predictive models based on strong features found in the data have higher accuracy.

Fourth, graphs provide a way to ensure transparency in AI decision-making, which makes the conclusions drawn through AI more interpretable. AI and machine learning have great application potential, and graphs unlock that potential. This is because graph database technology supports domain-related knowledge and linked data, making AI more widely applicable.

As far as the overall trend is concerned, we can foresee that in the era of big data, data loss is no longer the biggest challenge. What we are eager for is the ability to mine the value of data, and a large part of the value of data lies in the association between data. As the most effective technologies and methods for processing linked data, graph database and graph analysis will surely continue to shine and write a new chapter in database applications.

 The author of the database article in this issue of "New Programmer", thanks Mr. Yu for sharing

Inside the "New Programmer" magazine 

 Welcome more authors and readers to join "New Programmer"

Contact Invitation and Cooperation For bulk purchases, please contact yuzhong125

Guess you like

Origin blog.csdn.net/programmer_editor/article/details/124474982