Pethuru Raj (editor) - Demystifying Graph Data Science: Graph algorithms, analytics methods, platforms, databases, and use cases (Computing and Networks)

Srinivas Kumar Palvadi, K.G.M. Pradeep, Thota Siva Ratna Sai, D. Rammurthy and Vishal Dutt

Pethuru Raj, P. Beaulah Soundarabai and Peter Augustine

P. Divya, S. Jayalakshmi, M. Pavithra, R. Rajmohan, T. Ananth Kumar and Osamah Ibrahim Khalaf

Mats Agerstam

S. Usharani, K. Dhanalakshmi, P. Manju Bala, R. Rajmohan and S. Arunmozhi Selvi

P. Manju Bala, S. Usharani, T. Ananth Kumar, R. Rajmohan, M. Pavithra and G. Glorindal

N. Padmapriya, N. Kumaratharan, R. Rajmohan, T. Ananth Kumar, M. Pavithra and R. Dinesh Jackson Samuel

Vishal Dutt, Shweta Sharma and Swarn Avinash Kumar

N. Pooranam, S. Oswalt Manoj, G. Ignisha Rajathi and M. Amala Jayanthi

S. Thanga Ramya, V.P.G. Pushparathi, D. Praveena, A. Sumaiya Begum, B. Kalpana and Thangam Palani Swamy

Sajeev Ram Arumugam, Anusha Bamini, S. Oswalt Manoj, Rashmita Khilar and V. Felshi Sheeba

Neha Gupta and Rashmi Agrawal

Pethuru Raj and Nachamai Muthuraman

Yash Joshi and Siddhesh Kapote

Pethuru Raj and Nachamai Muthuraman

Pethuru Raj, D. Peter Augustine and P. Beaulah Soundarabai

Graph algorithms, analytics methods, platforms, databases, and use cases

Edited by
Pethuru Raj, Abhishek Kumar, Vicente Garca Daz and Nachamai Muthuraman

The Institution of Engineering and Technology

With the faster maturity and stability of digital technologies, we are being bombarded with zillions of digital entities, connected devices, and microservices. These interact purposefully to create huge sets of poly-structured digital data. The challenge is how to transition data into information and into knowledge. There are data analytics methods in plenty. The pace of data analytics is gaining the much-needed speed and sagacity with the continuous contributions of product and tool vendors. Data science is the domain increasingly associated with data analytics. There are big, fast and streaming data analytics platforms, frameworks, accelerators, toolkits, etc. for making data analytics simpler, faster and affordable.

In the big data world, NoSQL and distributed SQL databases gained the market and mind shares fast. Graph databases are one of the prominent NoSQL databases. Data representation through graphs has laid down a stimulating foundation to visualize and realize a stream of fresh capabilities.

On the other hand, the analytical competency is significantly improved through the faster maturity and stability of artificial intelligence (AI) algorithms [machine and deep learning (ML/DL)]. Thus, the classical and current data science paradigm is substantially advanced to have sophisticated abilities through the direct and distinct empowerment of AI algorithms. There is a twist now. Applying the AI-inspired data science methods on graph-structured data is being seen as a clear-cut gamechanger for the digital world. Extracting hidden patterns, useful associations, impending risks, future opportunities, and other useful and usable insights out of data heaps through data science platforms, frameworks, and engines is the new normal. Especially data science on graph data is acquiring special significance as there is a solid understanding that the blending of graphs and data science techniques can bring in a lot of noteworthy innovations and transformations.

Corporations are continuously seeking fresh ways to use their data to drive business innovations and disruptions to bring in real digital transformation. Supported by query languages, databases, algorithms, platforms, analytics methods and machine and deep learning (ML and DL) algorithms, graphs are now emerging as a new data structure for optimally representing a variety of data and their intimate relationships. This edited book aims to explain the various aspects and importance of graph data science. Graph analytics is being touted as the best way forward compared to traditional analytics methods. These methods are intrinsically capable of creating business value by intelligently leveraging connected data. The connectedness of data points facilitates the identification of clusters of related data points based on levels of influence, association, interaction frequency, and probability. Graph analytics is being empowered through a host of path-breaking analytics techniques to explore and pinpoint beneficial relationships between different entities such as organizations, people, transactions, etc.

The overwhelming idea is to create a database of things connecting to other things. Those things might be people connecting to other people through social and professional websites. Or they may be flights flying between cities across the globe. Graphs are hugely popular in enhancing search capabilities, recommending products to online buyers, detecting fraud, identifying the shortest route from one place to another, etc.

Graph data science is a technology-driven approach to discover knowledge from graph-represented data. Experts have pointed out that graph data science has the inherent strength to bring forth a suite of business, technical and user cases.

In this book, we are to cover the various aspects of graph data science and how it can be a game-changer for the data analytics domain. The prominent chapters of this book are

Graph technology This chapter covers graph theory concepts to build a strong and sustainable foundation to better understand graph analytics techniques and tools. Different types of graphs are discussed, as well as the latest trends and transitions happening in the graph technology space.

Depicting graph algorithms Having understood the significance of graph algorithms and analytics, researchers, and experts have come up with a number of graph-specific algorithms which are enabling and empowering graph analytics. This chapter presents promising and prominent graph algorithms such as Community Detection which detects group clustering or partition options; Centrality (Importance) which is for determining the significance of distinct nodes in the network; Similarity which evaluates how similar nodes are; Heuristic Link Prediction that estimates the likelihood of nodes forming a relationship; and Pathfinding & Search which finds optimal paths and evaluates route availability and quality.

Introducing graph analytics All kinds of collected and cleansed data are being investigated to extrapolate hidden insights out of data. Graph analytics is gaining prominence as traditional data analytics methods have failed to bring deeper and decisive insights out of data. Graph analytics (also called network analysis) is the analysis of relations among entities such as customers, products, solutions, services, operations, and devices. This chapter focuses on promising and potential techniques for augmenting graph analytics.

Graph databases and toolkits A graph database is a database designed to treat the relationships between data as equally important to the data itself. They are used for performing advanced graph analytics by connecting nodes and creating relationships (edges) in the form of graphs that can be queried by users. This chapter introduces leading graph databases and how they can simplify next-generation graph analytics.

Business use cases In this chapter, we present several business use cases showing applications of graph analytics such as clustering, partitioning, search, shortest path solution, widest path solution, finding connected components, and page rank.

Towards graph data science Machine and deep learning (ML/DL) algorithms support the discovery in real time of personalized, predictive, and prescriptive insights out of data via Big Data and Streaming analytics platforms. With the maturity of graph analytics, data science will get a strong boost from graph data science. This chapter shows data science capabilities and how the domain of data science is to be strengthened and solidified with the arrival of graph analytics.