TigerGraph releases version 3.0 of its graph database, asks techies for assistance with COVID-19 analytics

New features designed to enhance ease of use

TigerGraph, the company set up in 2012 by former Teradata and Twitter employee Yu Xu, has released version 3.0 of its distributed parallel graph database.

The new version contains a number of features designed to make the database easier to use by non-technical staff. The first of these is a wizard for importing data direct from a SQL database such as a datawarehouse or corporate CMS database into TigerGraph without the need for coding.

In a similar vein, TigerGraph 3.0 allows users to create and run graph queries via a drag-and-drop interface using a new tool called Visual Query Builder, again with a no-code approach.

"This allows non techies to be able to map out what they want to connect, then the computation and querying is done behind the scenes," said COO Todd Blaschka.

The third feature that's new to version 3 is user-defined indexing. This is a secondary index that helps the user jump directly to the data they want in order to speed up regularly run queries.

Blascha said the firm has also upped the speed of deployment for large-scale, multi-terabyte distributed graphs and improved the ease with which users can upgrade to a new version of the database.

TigerGraph has its own query language GSQL which is designed to make adoption by RDBMS users simpler, and is capable of both transactional and analytical processing.

Graph databases are growing in popularity as more organisations need to analyse disparate data from multiple sources. :Unlike relational and other types of NoSQL database, graph databases treat the relationships between datapoints as a first-class citizen. Common uses cases include anti-fraud systems, where you want to drill down and find connections between entities that are not obvious to the human eye. Another one is marketing 360, and telecoms firms use them to analyse phone usage.

One customer, Pagantis, uses TigerGraph to analyse a purchasers' credit rating in real-time as a transaction is taking place.

"One of the big things is uncovering hidden connections," Blaschka said "These are often two, three layers or four layers deep in the data and now I can find those patterns; that's really where the nuggets are."

A key attribute of graph databases is flexibility, since data does not have to be modelled as tables.

"You can modify the schema very quickly and it can be updated very easily, unlike with relational where you if want to do a schema update it's just such a big process," Blaschka said.

The company seeking technologists with skills in Python, Kafka and data wrangling to help discover insights into the coronavirus pandemic, to provide models for early warning systems, response efforts and impact assessment systems. The ​COVID19 Repository was started last month by TigerGraph developer Jonathan Herke.