Playing with Social Network, Geo-Spatial, Financial Flow, and Banking System Data: Graph-Theoretic Computation in Python Much of 'Big Data' revolution has to do with dealing with non-traditional datatypes and feeding machine learning engines with non-numerical variables. Just as Convolutional Neural Networks excel at object recognition task precisely because the convolution layers expressly capture grid-array information embedded in the 2-dimensional visual field, analysis involving network-patterned objects (individuals in an online social network, transportation nodes, banking accounts with payment network, etc.) stands to benefit from the ability to extract graph-theoretic properties from the network connectivity patterns, either to provide insights in and of themselves, or as pre-processing to further machine learning task down the line.
- (Bachelor's) Physics & Economics (double major), Case Western Reserve University
- (Master's) Operations Research (minor in Finance), Weatherhead School of Management
- (Doctorate's) Computational Intelligence & Operational Research, Imperial College, London
- (Bank of Thailand) Head of Quantitative Models & Financial Engineering, Financial Supervision Group
- (SCB - current) FSVP, Quantitative Models & Enterprise Analytics, Business Intelligence, Transformation Group