In this article, Dan McCreary, a distinguished engineer in artificial intelligence and graphs at Optum, deep dives into when large Fortune 10 healthcare companies should employ graph technologies. Initially, a checklist on what sort of challenges were a good fit for graphs was compiled based on his experience in a wide variety of domains. When this list was compared with a list of requirements where graphs do not perform well by themselves patterns began to emerge. Subsequently, methods for grouping similar use cases together were explored, as it would help better organize the growing list of business use cases.
Typically, grouping similar items—in this case use cases—together denote the beginning of building a classification system, an initial taxonomy of concepts. However, classification also can be very biased as it is will be based on the views of the group that is doing it. To eliminate this bias, data points were gathered by leveraging how a popular graph database classifies graph use cases. The differences between the two lists—one created to kick start the building of taxonomy of graph use cases and the other generated by the popular graph database—were apparent. The initial list was based on the key features that differentiate graph databases from relational databases, whereas, the use cases of the popular graph database were solution oriented.
In addition, the auto-complete list offered by search engines was also used, as it is moderately an objective way to classify graph usage. This helped generate high-level patterns such as those found below:
However, this taxonomy or classification cannot be deemed as completed. Some industry-specific use cases will draw from multiple areas of graph use case taxonomy. The more the matches better are the chances that a problem can be solved by a graph database.
Click here to read the article.
Please give your feedback on this article or share a similar story for publishing by clicking here.
