Presented By O'Reilly and Cloudera
Make Data Work
Dec 4–5, 2017: Training
Dec 5–7, 2017: Tutorials & Conference

Practical applications for graph techniques in supply chain analysis and finance

Eric Tham (National University of Singapore), Radha Pendyala (Thomson Reuters)
2:35pm3:15pm Wednesday, December 6, 2017
Average rating: *****
(5.00, 1 rating)

Who is this presentation for?

  • Data scientists, supply chain managers, and portfolio and risk managers

Prerequisite knowledge

  • A basic understanding of graphical analysis, GraphX, and NetworkX

What you'll learn

  • Understand the practical applications of graphical analysis


Graphical techniques are increasingly being used for big data. These techniques can be broadly classified into the three C’s: centrality, clustering, and connectedness. Eric Tham explains how these concepts are applied to supply chain analysis and financial portfolio management.

Centrality identifies the root causes of events or items in a graphical network—an example is the PageRank algorithm. Connectedness identifies relations between two or more items in the network—an example is Dijkstra’s algorithm. An example of clustering would be grouping items around similar themes with techniques like k-means clustering and affinity propagation.

In supply chain analysis, strategic prospects are clustered into hubs using affinity propagation. Hubs highlight symbiotic relationships among graphical nodes. Through the centrality measures, important nodes in the graphs are also strategically identified in the supply chain.

Graphical networks are also used in financial credit risk management, where research shows a most stable network is one with moderate connectedness. In financial portfolio management, a key criterion is the risk-reward payoff through diversification. This can be quantified through the use of graphical connectedness measures instead of the traditional correlation matrix.

Eric illustrates these concepts through commonly used open source graphical libraries like GraphX and NetworkX.

Photo of Eric Tham

Eric Tham

National University of Singapore

Eric Tham is an associate lecturer at the National University of Singapore. Previously, he was an enterprise data scientist at Thomson Reuters, led the quantitative data science team in a Chinese fintech startup with five million users, and worked in the financial industry in risk management, quantitative development, and energy economics with banks and oil companies. Over his career, he has developed sentiment indices from social media data and is an expert in unstructured data analysis, NLP, and machine learning in financial applications. He is a frequent speaker at conferences and contributed a chapter to the Handbook of Sentiment Analysis in Finance.

Radha Pendyala

Thomson Reuters

Radha works as an Enterprise Data Scientist at Thomson Reuters. His work involves applying machine learning and quantitative financial modeling techniques to large datasets in order
to solve specific problems in the financial sector. Prior to Reuters, he has worked as a
portfolio manager at Goldman Sachs Asset Management. He has more than a decade of
experience in building financial and statistical models.
Radha has obtained his masters in financial engineering degree from City University of
New York, a post graduate degree in management from Indian Institute of Management,
Indore and a B.Tech in Civil Engineering from Indian Institute of Technology, Madras.