Background & approach

By H2 2017, Twitter already had a wealth of econometric and conventional survey-based evidence to prove the platform's effectiveness. What was needed was something fresh and new to articulate how Twitter works to spread information and how advertisers can best leverage the platform for communication.

Our story propagation study has generated the first market-facing information propagation model of its kind – with help from IPG Mediabrands' Data team. The model is based not merely on volume, but on a range of other metrics and combines big data analytics with human-led meta-analysis in a truly innovative way.

It tackled the question of information propagation at scale. To deliver this project, Twitter & Mediabrands' Data Team were required to build a bespoke data engine utilizing a mix of programming languages such as R, JavaScript and HTML to scrape and scrutinize the huge dataset. This approach, build bespoke for Twitter, has the ability to capture story propagation across any networked dataset, and can capture granular activities across comment sections, blogging platforms and social medias.

Methodology