Here's a few of the fascinating projects we've been privileged to participate in lately.

ADAPTIVE LEARNING models, 3P Learning

3P Learning are the creators of Mathletics educational software, used by schools in 61 countries to supplement their Math curriculum with fun interactive math activities. 3P wanted to move from simple adaptive learning at classroom level to full personalisation of students' learning programs based on real time performance on activities and assessments. The challenge? Over 100 different curriculums, 7 million students and very large dynamic data sets with tables up to 780,000,000 records to be processed. We worked with 3P and their technology partners to develop a scalable personalisation model using highly segmented collaborative filtering algorithms, deployed across a distributed architecture for maximum speed. A great challenging project, with a feel-good factor - we're helping the next generation of data scientists learn to love math! 

what makes a party rock? coca cola  

Coca-Cola's brand has always been social and about fun times with friends. For Summer 2016/17, Coca Cola Australia's  experiential agency, Urban, came up with a quirky plan to create 'Coke Moments' and offer consumers free product to create their own CocaCola fuelled party. We used text mining and sentiment analysis on two years of Twitter data to identify the keywords and emotions co-associated to CocaCola and summer parties. We then worked with Urban to set up a Twitter call to action for "Coca Cola Moments' and live monitoring of consumer tweets so that they could deploy their field crew to deliver Coca Cola product to parties as they happened. 

why gen z don't want your cds, warner music

Like all in the music business, Warner Music is working through how to maintain their relevance in an age of digital downloads and music streaming. Warners' digital team had carried out a number of surveys on their various mailing lists and digital groups, but had not had a chance to do anything with the data beyond some high level reporting. The team wanted to get into the minds of different age cohorts and understand their relationship with music - genres, consumption methods, buying vs streaming - so as to create more relevant engagement. We compiled and analysed their survey data, and produced an insights report. Not surprisingly, the Gen Z tweens and teens have a totally different relationship to music, listening and brands than their older counterparts!