Zafar Gilani

Stweeler: A Framework for Twitter Bot Analysis

The WWW has seen a massive growth in variety and usage of OSNs. The rising population of users on Twitter and its open nature has made it an ideal platform for various kinds of opportunistic pursuits, such as news and emergency communication, business promotion, political campaigning, spamming and spreading malicious content. Most of these opportunistic pursuits are exploited through automated programs, known as bots. In this study we propose a framework (Stweeler) to study bot impact and influence on Twitter from systems and social media perspectives.

Do Bots impact Twitter activity

The WWW has seen massive growth in population of automated programs (bots) for a variety of exploits on online social networks (OSNs). In this paper we extend on our previous work to study the affects of bots on Twitter. By setting up a bot account on Twitter and conducting analysis on a click logs dataset from our web server, we show that despite bots being in smaller numbers, they exercise a profound impact on content popularity and activity on Twitter.

Classification of Twitter Accounts into Automated Agents and Human Users

Online social networks (OSNs) have seen a remarkable rise in the presence of surreptitious automated accounts. Massive human user-base and business-supportive operating model of social networks (such as Twitter) facilitates the creation of automated agents. In this paper we outline a systematic methodology and train a classifier to categorise Twitter accounts into ‘automated’ and ‘human’ users. To improve classification accuracy we employ a set of novel steps. First, we divide the dataset into four popularity bands to compensate for differences in types of accounts. Second, we create a large ground truth dataset using human annotations and extract relevant features from raw tweets. To judge accuracy of the procedure we calculate agreement among human annotators as well as with a bot detection research tool. We then apply a Random Forests classifier that achieves an accuracy close to human agreement. Finally, as a concluding step we perform tests to measure the efficacy of our results.

The Implications of Twitterbot Generated Data Traffic on Networked Systems

The explosion of bots on the Web brings an unprecedented increase in traffic from non-human sources. This work studies bot traffic on Twitter, finding that almost 50% of traffic is generated and propagated by a rapidly growing bot population -- a major concern for networked systems in the future.

Inferring network infrastructural behaviour during disasters

An unexpected increase in natural disasters has prompted a large interest in governments and organisations to utilise ICT for many different purposes such as preparation, impact mitigation, loss reduction and relief efforts. This paper presents initial work on studying disaster scenarios from device level perspective to characterise network infrastructural behaviour during extraordinary situations. We find connectivity challenges during disasters and observe sharp decline of quality metrics and loss of station quantity between ordinary and extraordinary time periods. We also make distinctions between usual and unusual behaviour seen during ordinary and extraordinary situations.

Peer provided cell-like networks built out of thin air

The success of Wi-Fi technology as an efficient and low-cost last-mile access solution has enabled massive spontaneous deployments generating storms of beacons all across the globe. Emerging location systems are using these beacons to observe mobility patterns of people through portable or wearable devices and offer use-cases that can help solve critical problems in the developing world. In this paper, we design and develop a novel prototype to organise these spontaneous deployments of Access Points into what we call virtual cells (vcells). We compute virtual cells from a list of Access Points collected from different active scans for a geographical region. We argue that virtual cells can be encoded using Bloom filters to implement the location process. Lastly, we present two illustrative use-cases to showcase the suitability and challenges of the technique.