Data mining with LinkedIn

Abstract

Scott McClellan didn't have a very good May. The chief technologist and interim vice-president of engineering for HP's fledgling cloud services business had updated his LinkedIn profile. In the update, he described his work in rolling out a future cloud-based service that the company would implement. Unfortunately, HP hadn't yet announced it, leading to press reports that detailed every word of his unfortunate error.1 The information soon disappeared from his profile, doubtless pulled at the behest of his employer. LinkedIn is a valuable source of business network information. It is also a way of enumerating networks of individuals at a micro and a macro scale. The social networking service is a perfect example of ‘big data’ – a very large dataset that cannot be mined using traditional relational database management tools. However, it is relatively easy to analyse small subsets (such as personal networks of contacts) for potentially useful results. By exploiting profile information and the linkages provided by connections made by individual users and the groups function, it is possible to map an individual's interests and network of contacts. It's also possible to enumerate entire organisations. LinkedIn also provides an API to which it's possible to connect using a variety of data mining and visualisation tools. All this information can be used for commercial purposes or by malicious people looking for hacking or social engineering targets, explains Danny Bradbury.