451logo “We are seeing increasing interest in graph database technologies…as a platform for new analytic applications that cannot be supported by traditional data models.”
Cray's YarcData claims early success for graph database appliance (Matt Aslett, August 10, 2012)

Graph Analytics Challenge

YarcData Support

In December, YarcData, announced the six finalists for the YarcData Graph Analytics Challenge. The top six entries for the contest, which features $100,000 in prizes including a $70,000 grand prize for the first place winner, were determined to have entered the best submissions for un-partitionable, Big Data graph problems. The top six submissions span a number of diverse categories, such as healthcare, law enforcement, sports, life sciences and social media.

Analysis of multiple types of cancers and thousands of patients to identify potentially potent approved drugs that could be repurposed for a given cancer. This identification is based on multiple sources of information, including large-scale molecular measurements from The Cancer Genome Atlas, the MEDLINE database that includes millions of biomedical scholarly articles, as well as drug, disease, and molecular interactions databases. Performing queries on the Urika system resulted in many repurposing hits that warrant further experimental validation. (Dr. Brady Bernard, Dr. Ilya Shmulevich, Andrea Eakin)
Autism in its various forms is not well understood and is believed to be caused by a multitude of factors whose contribution may be different from case to case. A team lead by Adam Lugowski investigated causes of autism by understanding the clusters of precursors among diagnosed individuals. “We discovered that Urika really shines on a variety of queries as compared to COTS servers. This has allowed us to cluster very large RDF graphs, something not possible on lesser hardware,” said Adam Lugowski. (Adam Lugowski, Dr. John Gilbert, Kevin Deweese)
Improving the speed and accuracy of risk prediction methods, even by a small amount could have a major impact on individual and population health, as well as on health system costs and performance. Dr. Flaxman modeled mortality in the 30 days following a heart attack so that populations at particular risk can be identified. He developed and evaluated an approach to risk prediction particularly suited to big data analysis with the Urika graph appliance. (Dr. Abraham Flaxman)

Finalists for the YarcData Graph Analytics Challenge

finalist_iconThe author of Diamond Dollars and consultant to major league baseball teams, Vince Gennaro, analyzed individual pitch and batted ball data to determine how batters perform against specific types of pitchers. By clustering pitchers by similarity and measuring batting performance against pitcher clusters he was able to predict batter-pitcher outcomes against individual pitchers. This approached has opened up the door for the use of graph analytics in baseball and other professional sports, including the areas of predicting player performance, roster optimization, and contract valuation.
(Vince Genarro)

finalist_iconIn studying trust-based relationships, we can better understand the multi-level complexities of trust and how it relates to social phenomenon like mentoring, expertise, clandestine behaviors, etc. This team conducted large-scale social network analysis to determine the impact of team behavior/ collaboration by using gaming data. Many of the queries that were asked would not be possible on a distributed memory architecture, as there is no real way to partition the data. (Dominic DiFranzo, Bassem Makmi, Qingpeng Zhang)