Inverse network sampling to explore online brand allegiance

Abstract

Within the online media universe there are many underlying communities. These may be defined, for example, through politics, location, health, occupation, extracurricular interests or retail habits. Government departments, charities and commercial organisations can benefit greatly from insights about the structure of these communities; the move to customer-centered practices requires knowledge of the customer base. Motivated by this issue, we address the fundamental question of whether a subnetwork looks like a collection of individuals who have effectively been picked at random from the whole, or instead forms a distinctive community with a new, discernible structure. In the former case, to spread a message to the intended user base it may be best to use traditional broadcast media (TV, billboard), whereas in the latter case a more targeted approach could be more effective. In this work, we therefore formalize a concept of testing for substructure and apply it to social interaction data. First, we develop a statistical test to determine whether a given subnetwork (induced subgraph) is likely to have been generated by sampling nodes from the full network uniformly at random. This tackles an interesting inverse alternative to the more widely studied “forward” problem. We then apply the test to a Twitter reciprocated mentions network where a range of brand name based subnetworks are created via tweet content. We correlate the computed results against the independent views of sixteen digital marketing professionals. We conclude that there is great potential for social media based analytics to quantify, compare and interpret on-line brand allegiances systematically, in real time and at large scale

Network analysis and modelling: Special issue ofEuropean Journal of Applied Mathematics

“…In addition to discussing a fascinating application, this paper considers several important issues in networks, such as data sampling (which is ubiquitous in the study of networks), time-dependent interactions, and spatial constraints. The importance of sampling in the study of networks is also examined in our penultimate paper, by Grindrod et al [14], which takes an inverse-problem approach and studies an application to online brand allegiance in Twitter. Another central issue in networks is the study of important nodes, edges, and other substructures [23].…”

Beyond privacy and exposure: ethical issues within citizen-facing analytics

“…In such cases, the discovery of what works and some knowledge of why such methods can work has allowed companies to cease the indiscriminate broadcasting and move towards more personal marketing propositions. This in itself raises many interesting questions including which products and brands should be targeted into social networks, for example, and will generate buzz; and which products are inappropriate for this and need to relay on broadcasting (on the backs of buses, for example) [1,47]. Here the data exploited are usually proprietary data and its security is paramount (for reputational and shareholder value reasons for the corporate, and for privacy reason for the citizen-interests are highly aligned).…”