foreAction data analytics: Optimization of SQL and NoSQL hybrid architectures

点击翻译按钮获取中文摘要

A software development and technology consulting firm based in Ottawa has developed a software product for the political and non-profit space that collects, stores and analyzes contact data. Based on information stored within the data, the software is used to "micro-target" contacts for specific marketing and informational materials. The software is being developed to allow for the creation of contact lists along three specific axis: the likelihood of a customer to be responsive to a message (propensity score), the value/priority of the contact (value score), and the desired method of contact (medium score). The analytics of these three axis will determine whether a specific engagement must be initiated with the contact during a campaign of any nature. Data is collected continually via ongoing interactions with contacts, via social media polling, via third party statistics (demographics, polling), via surveys/petitions, at events, etc. There is a pressing need to balance the analytics to create valuable, targeted contact lists and a Business Intelligence dashboard (Key Performance Indicators) in a timely and effective way against the need for real time access to updates to the information. In the political space in particular, in the days leading up to an election, the volume of new data can be significant, and needs to be parsed and included in analytics as a part of the outreach strategy for voters. The proposed research looks into optimization techniques, algorithms (especially) and architectures that will allow to update data in real time, and still perform large-scale analytics on the data. An implementation of the solution developed during the research will be done for the consulting firm.