If you have not heard of graph databases, then you better catch up. Graph databases are one of the fastest growing areas in technology and their value is not well understood.
According to Gartner, graph database management systems will grow at 100% through 2022 to enable more complex data science.
The graph database market is expected to reach $2.4 billion by 2023, growing at 24% annually.
What is a database?
A database is a row-column structure. The columns (generally) represent the varying attributes and the rows represent the (homogeneous) data elements or records.
A database allows sorting, to show the ranking of various records in the database. Obviously, each record is different, even though they are all similar.
A database cannot show the relative weightage or importance of one record vs another. It also cannot show how the records are related to one another.
What is a graph database?
Unlike a traditional or a relational database, a graph database contains nodes, edges and properties to help understand the data elements better.
It easily shows the elements with higher weightage, or higher influencers and how they are related to other elements in the same database.
Graph databases become even more powerful when combined with other databases, such as products with sales transactions.
Applications of graph databases
The page at https://neo4j.com/use-cases/ provides a good listing of use cases where graph databases provide real value.
I thought the following three are easy to understand:
Master data.
At face value, master data may be the least intuitive to apply from a graph database perspective. But when you combine master data with the related transactions, e.g. product master data with sales transactions, then the application becomes very powerful. As an example, the graphical representation of product master can give you ideas on how to bundle products and provide discounted pricing, to maximize sales.
Influencers.
Whether on social media or within an enterprise, understanding the bigger influencers (and engaging them strategically) can help improve results. Applications could be in HR (finding influential recruiters) or restaurant promoters (influencers in social media channels) or even physicians as Key Opinion Leaders (to promote a company’s drugs).
Attribution.
In a long cycle, there would be a number of “nodes”, each with varying contributions towards the end-result. Based on the result, a graphical representation of the various nodes can help one understand the levels of contribution of each node towards the end-result. If in a supply chain, the product, at the end, turned out to be damaged, we can tell which node had a greater contribution towards the end-result.
Alternatively, in a long sales cycle, we can highlight the part that most contributed to the sale or the one that thwarted the sale.
Prognosis
The accelerated growth of graph databases notwithstanding, I believe that every database will (or must) have its graphical equivalent – in a couple years. Without the graphical representation and the valuable insights that can be gleaned, databases will just be a listing.
Recent Comments