在製造消費性電子上使用BigQuery的初步研究

Abstract

None

It is well known that several business sectors (advertising, insurance, finance, etc.) have been benefited from the big data revolution. However, in the manufacturing field the investment in it is relatively inactive. Previous research had pointed out that the barrier to bring big data into practice is usually facing barrier of inexpert stuff, cost and might cause a privation of sponsorship. Also, previous report estimates a lower potential value behind the data in manufacturing, however, it is hard to justify the value until someone hit the jackpot. In addition, looking into the portion that manufacturing occupied in the GDP, we believe it worths to make some tradeoffs and design a system for preliminary study that mainly optimized to be affordable in terms of infrastructure, skill set training and deployment without tumbling existing system in the field. In this thesis, we studied an approach backed by Dremel (Commercialized product name as BigQuery), a nested and column-based OLAP system. We focus on consumer electronics products, more specifically: laptop, tablet and router. Our preliminary result shows useful cases can be obtained immediately by collecting more data (we estimate at least two orders of magnitude in terms of data size) without using advanced data mining techniques.