Would you like data with that? Library provision of financial datasets

Within the QUT Business School (QUTBS)– researchers across economics, finance and accounting depend on data driven research. They analyze historic and global financial data across a range of instruments to understand the relationships and effects between them as they respond to news and events in their region. Scholars and Higher Degree Research Students in turn seek out universities which offer these particular datasets to further their research. This involves downloading and manipulating large datasets, often with a focus on depth of detail, frequency and long tail historical data. This is stock exchange data and has potential commercial value therefore the license for access tends to be very expensive. This poster reports the following findings: •The library has a part to play in freeing up researchers from the burden of negotiating subscriptions, fundraising and managing the legal requirements around license and access. •The role of the library is to communicate the nature and potential of these complex resources across the university to disciplines as diverse as Mathematics, Health, Information Systems and Creative Industries. •Has demonstrated clear concrete support for research by QUT Library and built relationships into faculty. It has made data available to all researchers and attracted new HDRs. The aim is to reach the output threshold of research outputs to submit into FOR Code 1502 (Banking, Finance and Investment) for ERA 2015. •It is difficult to identify what subset of dataset will be obtained given somewhat vague price tiers. •The integrity of data is variable as it is limited by the way it is collected, this occasionally raises issues for researchers(Cook, Campbell, & Kelly, 2012) •Improved library understanding of the content of our products and the nature of financial based research is a necessary part of the service.

R1SVM: A randomised nonlinear approach to large-scale anomaly detection

The problem of unsupervised anomaly detection arises in a wide variety of practical applications. While one-class support vector machines have demonstrated their effectiveness as an anomaly detection technique, their ability to model large datasets is limited due to their memory and time complexity for training. To address this issue for supervised learning of kernel machines, there has been growing interest in random projection methods as an alternative to the computationally expensive problems of kernel matrix construction and sup-port vector optimisation. In this paper we leverage the theory of nonlinear random projections and propose the Randomised One-class SVM (R1SVM), which is an efficient and scalable anomaly detection technique that can be trained on large-scale datasets. Our empirical analysis on several real-life and synthetic datasets shows that our randomised 1SVM algorithm achieves comparable or better accuracy to deep auto encoder and traditional kernelised approaches for anomaly detection, while being approximately 100 times faster in training and testing.

The role of informal networks in providing effective work opportunities for people with an intellectual disability

Being in paid employment is socially valued, and is linked to health, financial security and time use. Issues arising from a lack of occupational choice and control, and from diminished role partnerships are particularly problematic in the lives of people with an intellectual disability. Informal support networks are shown to influence work opportunities for people without disabilities, but their impact on the work experiences of people with disability has not been thoroughly explored. The experience of 'work' and preparation for work was explored with a group of four people with an intellectual disability (the participants) and the key members of their informal support networks (network members) in New South Wales, Australia. Network members and participants were interviewed and participant observations of work and other activities were undertaken. Data analysis included open, conceptual and thematic coding. Data analysis software assisted in managing the large datasets across multiple team members. The insight and actions of network members created and sustained the employment and support opportunities that effectively matched the needs and interests of the participants. Recommendations for future research are outlined.

Using Bayesian methods for the estimation of uncertainty in complex statistical models

The research objectives of this thesis were to contribute to Bayesian statistical methodology by contributing to risk assessment statistical methodology, and to spatial and spatio-temporal methodology, by modelling error structures using complex hierarchical models. Specifically, I hoped to consider two applied areas, and use these applications as a springboard for developing new statistical methods as well as undertaking analyses which might give answers to particular applied questions. Thus, this thesis considers a series of models, firstly in the context of risk assessments for recycled water, and secondly in the context of water usage by crops. The research objective was to model error structures using hierarchical models in two problems, namely risk assessment analyses for wastewater, and secondly, in a four dimensional dataset, assessing differences between cropping systems over time and over three spatial dimensions. The aim was to use the simplicity and insight afforded by Bayesian networks to develop appropriate models for risk scenarios, and again to use Bayesian hierarchical models to explore the necessarily complex modelling of four dimensional agricultural data. The specific objectives of the research were to develop a method for the calculation of credible intervals for the point estimates of Bayesian networks; to develop a model structure to incorporate all the experimental uncertainty associated with various constants thereby allowing the calculation of more credible credible intervals for a risk assessment; to model a single day’s data from the agricultural dataset which satisfactorily captured the complexities of the data; to build a model for several days’ data, in order to consider how the full data might be modelled; and finally to build a model for the full four dimensional dataset and to consider the timevarying nature of the contrast of interest, having satisfactorily accounted for possible spatial and temporal autocorrelations. This work forms five papers, two of which have been published, with two submitted, and the final paper still in draft. The first two objectives were met by recasting the risk assessments as directed, acyclic graphs (DAGs). In the first case, we elicited uncertainty for the conditional probabilities needed by the Bayesian net, incorporated these into a corresponding DAG, and used Markov chain Monte Carlo (MCMC) to find credible intervals, for all the scenarios and outcomes of interest. In the second case, we incorporated the experimental data underlying the risk assessment constants into the DAG, and also treated some of that data as needing to be modelled as an ‘errors-invariables’ problem [Fuller, 1987]. This illustrated a simple method for the incorporation of experimental error into risk assessments. In considering one day of the three-dimensional agricultural data, it became clear that geostatistical models or conditional autoregressive (CAR) models over the three dimensions were not the best way to approach the data. Instead CAR models are used with neighbours only in the same depth layer. This gave flexibility to the model, allowing both the spatially structured and non-structured variances to differ at all depths. We call this model the CAR layered model. Given the experimental design, the fixed part of the model could have been modelled as a set of means by treatment and by depth, but doing so allows little insight into how the treatment effects vary with depth. Hence, a number of essentially non-parametric approaches were taken to see the effects of depth on treatment, with the model of choice incorporating an errors-in-variables approach for depth in addition to a non-parametric smooth. The statistical contribution here was the introduction of the CAR layered model, the applied contribution the analysis of moisture over depth and estimation of the contrast of interest together with its credible intervals. These models were fitted using WinBUGS [Lunn et al., 2000]. The work in the fifth paper deals with the fact that with large datasets, the use of WinBUGS becomes more problematic because of its highly correlated term by term updating. In this work, we introduce a Gibbs sampler with block updating for the CAR layered model. The Gibbs sampler was implemented by Chris Strickland using pyMCMC [Strickland, 2010]. This framework is then used to consider five days data, and we show that moisture in the soil for all the various treatments reaches levels particular to each treatment at a depth of 200 cm and thereafter stays constant, albeit with increasing variances with depth. In an analysis across three spatial dimensions and across time, there are many interactions of time and the spatial dimensions to be considered. Hence, we chose to use a daily model and to repeat the analysis at all time points, effectively creating an interaction model of time by the daily model. Such an approach allows great flexibility. However, this approach does not allow insight into the way in which the parameter of interest varies over time. Hence, a two-stage approach was also used, with estimates from the first-stage being analysed as a set of time series. We see this spatio-temporal interaction model as being a useful approach to data measured across three spatial dimensions and time, since it does not assume additivity of the random spatial or temporal effects.

A Bayesian analysis of an agricultural field trial with three spatial dimensions

Modern technology now has the ability to generate large datasets over space and time. Such data typically exhibit high autocorrelations over all dimensions. The field trial data motivating the methods of this paper were collected to examine the behaviour of traditional cropping and to determine a cropping system which could maximise water use for grain production while minimising leakage below the crop root zone. They consist of moisture measurements made at 15 depths across 3 rows and 18 columns, in the lattice framework of an agricultural field. Bayesian conditional autoregressive (CAR) models are used to account for local site correlations. Conditional autoregressive models have not been widely used in analyses of agricultural data. This paper serves to illustrate the usefulness of these models in this field, along with the ease of implementation in WinBUGS, a freely available software package. The innovation is the fitting of separate conditional autoregressive models for each depth layer, the ‘layered CAR model’, while simultaneously estimating depth profile functions for each site treatment. Modelling interest also lay in how best to model the treatment effect depth profiles, and in the choice of neighbourhood structure for the spatial autocorrelation model. The favoured model fitted the treatment effects as splines over depth, and treated depth, the basis for the regression model, as measured with error, while fitting CAR neighbourhood models by depth layer. It is hierarchical, with separate onditional autoregressive spatial variance components at each depth, and the fixed terms which involve an errors-in-measurement model treat depth errors as interval-censored measurement error. The Bayesian framework permits transparent specification and easy comparison of the various complex models compared.

Relaxed phylogenetics and dating with confidence

In phylogenetics, the unrooted model of phylogeny and the strict molecular clock model are two extremes of a continuum. Despite their dominance in phylogenetic inference, it is evident that both are biologically unrealistic and that the real evolutionary process lies between these two extremes. Fortunately, intermediate models employing relaxed molecular clocks have been described. These models open the gate to a new field of “relaxed phylogenetics.” Here we introduce a new approach to performing relaxed phylogenetic analysis. We describe how it can be used to estimate phylogenies and divergence times in the face of uncertainty in evolutionary rates and calibration times. Our approach also provides a means for measuring the clocklikeness of datasets and comparing this measure between different genes and phylogenies. We find no significant rate autocorrelation among branches in three large datasets, suggesting that autocorrelated models are not necessarily suitable for these data. In addition, we place these datasets on the continuum of clocklikeness between a strict molecular clock and the alternative unrooted extreme. Finally, we present analyses of 102 bacterial, 106 yeast, 61 plant, 99 metazoan, and 500 primate alignments. From these we conclude that our method is phylogenetically more accurate and precise than the traditional unrooted model while adding the ability to infer a timescale to evolution.

Prediction of transcriptional regulatory interactions in bacteria : a comparative genomics approach

Exponential growth of genomic data in the last two decades has made manual analyses impractical for all but trial studies. As genomic analyses have become more sophisticated, and move toward comparisons across large datasets, computational approaches have become essential. One of the most important biological questions is to understand the mechanisms underlying gene regulation. Genetic regulation is commonly investigated and modelled through the use of transcriptional regulatory network (TRN) structures. These model the regulatory interactions between two key components: transcription factors (TFs) and the target genes (TGs) they regulate. Transcriptional regulatory networks have proven to be invaluable scientific tools in Bioinformatics. When used in conjunction with comparative genomics, they have provided substantial insights into the evolution of regulatory interactions. Current approaches to regulatory network inference, however, omit two additional key entities: promoters and transcription factor binding sites (TFBSs). In this study, we attempted to explore the relationships among these regulatory components in bacteria. Our primary goal was to identify relationships that can assist in reducing the high false positive rates associated with transcription factor binding site predictions and thereupon enhance the reliability of the inferred transcription regulatory networks. In our preliminary exploration of relationships between the key regulatory components in Escherichia coli transcription, we discovered a number of potentially useful features. The combination of location score and sequence dissimilarity scores increased de novo binding site prediction accuracy by 13.6%. Another important observation made was with regards to the relationship between transcription factors grouped by their regulatory role and corresponding promoter strength. Our study of E.coli ��70 promoters, found support at the 0.1 significance level for our hypothesis | that weak promoters are preferentially associated with activator binding sites to enhance gene expression, whilst strong promoters have more repressor binding sites to repress or inhibit gene transcription. Although the observations were specific to �70, they nevertheless strongly encourage additional investigations when more experimentally confirmed data are available. In our preliminary exploration of relationships between the key regulatory components in E.coli transcription, we discovered a number of potentially useful features { some of which proved successful in reducing the number of false positives when applied to re-evaluate binding site predictions. Of chief interest was the relationship observed between promoter strength and TFs with respect to their regulatory role. Based on the common assumption, where promoter homology positively correlates with transcription rate, we hypothesised that weak promoters would have more transcription factors that enhance gene expression, whilst strong promoters would have more repressor binding sites. The t-tests assessed for E.coli �70 promoters returned a p-value of 0.072, which at 0.1 significance level suggested support for our (alternative) hypothesis; albeit this trend may only be present for promoters where corresponding TFBSs are either all repressors or all activators. Nevertheless, such suggestive results strongly encourage additional investigations when more experimentally confirmed data will become available. Much of the remainder of the thesis concerns a machine learning study of binding site prediction, using the SVM and kernel methods, principally the spectrum kernel. Spectrum kernels have been successfully applied in previous studies of protein classification [91, 92], as well as the related problem of promoter predictions [59], and we have here successfully applied the technique to refining TFBS predictions. The advantages provided by the SVM classifier were best seen in `moderately'-conserved transcription factor binding sites as represented by our E.coli CRP case study. Inclusion of additional position feature attributes further increased accuracy by 9.1% but more notable was the considerable decrease in false positive rate from 0.8 to 0.5 while retaining 0.9 sensitivity. Improved prediction of transcription factor binding sites is in turn extremely valuable in improving inference of regulatory relationships, a problem notoriously prone to false positive predictions. Here, the number of false regulatory interactions inferred using the conventional two-component model was substantially reduced when we integrated de novo transcription factor binding site predictions as an additional criterion for acceptance in a case study of inference in the Fur regulon. This initial work was extended to a comparative study of the iron regulatory system across 20 Yersinia strains. This work revealed interesting, strain-specific difierences, especially between pathogenic and non-pathogenic strains. Such difierences were made clear through interactive visualisations using the TRNDifi software developed as part of this work, and would have remained undetected using conventional methods. This approach led to the nomination of the Yfe iron-uptake system as a candidate for further wet-lab experimentation due to its potential active functionality in non-pathogens and its known participation in full virulence of the bubonic plague strain. Building on this work, we introduced novel structures we have labelled as `regulatory trees', inspired by the phylogenetic tree concept. Instead of using gene or protein sequence similarity, the regulatory trees were constructed based on the number of similar regulatory interactions. While the common phylogentic trees convey information regarding changes in gene repertoire, which we might regard being analogous to `hardware', the regulatory tree informs us of the changes in regulatory circuitry, in some respects analogous to `software'. In this context, we explored the `pan-regulatory network' for the Fur system, the entire set of regulatory interactions found for the Fur transcription factor across a group of genomes. In the pan-regulatory network, emphasis is placed on how the regulatory network for each target genome is inferred from multiple sources instead of a single source, as is the common approach. The benefit of using multiple reference networks, is a more comprehensive survey of the relationships, and increased confidence in the regulatory interactions predicted. In the present study, we distinguish between relationships found across the full set of genomes as the `core-regulatory-set', and interactions found only in a subset of genomes explored as the `sub-regulatory-set'. We found nine Fur target gene clusters present across the four genomes studied, this core set potentially identifying basic regulatory processes essential for survival. Species level difierences are seen at the sub-regulatory-set level; for example the known virulence factors, YbtA and PchR were found in Y.pestis and P.aerguinosa respectively, but were not present in both E.coli and B.subtilis. Such factors and the iron-uptake systems they regulate, are ideal candidates for wet-lab investigation to determine whether or not they are pathogenic specific. In this study, we employed a broad range of approaches to address our goals and assessed these methods using the Fur regulon as our initial case study. We identified a set of promising feature attributes; demonstrated their success in increasing transcription factor binding site prediction specificity while retaining sensitivity, and showed the importance of binding site predictions in enhancing the reliability of regulatory interaction inferences. Most importantly, these outcomes led to the introduction of a range of visualisations and techniques, which are applicable across the entire bacterial spectrum and can be utilised in studies beyond the understanding of transcriptional regulatory networks.

Talking of many things : using topical networks to study discussions in social media

This paper outlines a method for studying online activity using both qualitative and quantitative methods: topical network analysis. A topical network refers to "the collection of sites commenting on a particular event or issue, and the links between them" (Highfield, Kirchhoff, & Nicolai, 2011, p. 341). The approach is a complement for the analysis of large datasets enabling the examination and comparison of different discussions as a means of improving our understanding of the uses of social media and other forms of online communication. Developed for an analysis of political blogging, the method also has wider applications for other social media websites such as Twitter.

A polymorphism in the HLA-DPB1 gene is associated with susceptibility to multiple sclerosis

We conducted an association study across the human leukocyte antigen (HLA) complex to identify loci associated with multiple sclerosis (MS). Comparing 1927 SNPs in 1618 MS cases and 3413 controls of European ancestry, we identified seven SNPs that were independently associated with MS conditional on the others (each ). All associations were significant in an independent replication cohort of 2212 cases and 2251 controls () and were highly significant in the combined dataset (). The associated SNPs included proxies for HLA-DRB1*15:01 and HLA-DRB1*03:01, and SNPs in moderate linkage disequilibrium (LD) with HLA-A*02:01, HLA-DRB1*04:01 and HLA-DRB1*13:03. We also found a strong association with rs9277535 in the class II gene HLA-DPB1 (discovery set , replication set , combined ). HLA-DPB1 is located centromeric of the more commonly typed class II genes HLA-DRB1, -DQA1 and -DQB1. It is separated from these genes by a recombination hotspot, and the association is not affected by conditioning on genotypes at DRB1, DQA1 and DQB1. Hence rs9277535 represents an independent MS-susceptibility locus of genome-wide significance. It is correlated with the HLA-DPB1*03:01 allele, which has been implicated previously in MS in smaller studies. Further genotyping in large datasets is required to confirm and resolve this association.

Metrics for Understanding Communication on Twitter

As the systematic investigation of Twitter as a communications platform continues, the question of developing reliable comparative metrics for the evaluation of public, communicative phenomena on Twitter becomes paramount. What is necessary here is the establishment of an accepted standard for the quantitative description of user activities on Twitter. This needs to be flexible enough in order to be applied to a wide range of communicative situations, such as the evaluation of individual users’ and groups of users’ Twitter communication strategies, the examination of communicative patterns within hashtags and other identifiable ad hoc publics on Twitter (Bruns & Burgess, 2011), and even the analysis of very large datasets of everyday interactions on the platform. By providing a framework for quantitative analysis on Twitter communication, researchers in different areas (e.g., communication studies, sociology, information systems) are enabled to adapt methodological approaches and to conduct analyses on their own. Besides general findings about communication structure on Twitter, large amounts of data might be used to better understand issues or events retrospectively, detect issues or events in an early stage, or even to predict certain real-world developments (e.g., election results; cf. Tumasjan, Sprenger, Sandner, & Welpe, 2010, for an early attempt to do so).

Tensor-based item recommendation using probabilistic ranking in social tagging systems

A common problem with the use of tensor modeling in generating quality recommendations for large datasets is scalability. In this paper, we propose the Tensor-based Recommendation using Probabilistic Ranking method that generates the reconstructed tensor using block-striped parallel matrix multiplication and then probabilistically calculates the preferences of user to rank the recommended items. Empirical analysis on two real-world datasets shows that the proposed method is scalable for large tensor datasets and is able to outperform the benchmarking methods in terms of accuracy.

A tag-based personalized item recommendation system using tensor modeling and topic model approaches

This research falls in the area of enhancing the quality of tag-based item recommendation systems. It aims to achieve this by employing a multi-dimensional user profile approach and by analyzing the semantic aspects of tags. Tag-based recommender systems have two characteristics that need to be carefully studied in order to build a reliable system. Firstly, the multi-dimensional correlation, called as tag assignment , should be appropriately modelled in order to create the user profiles [1]. Secondly, the semantics behind the tags should be considered properly as the flexibility with their design can cause semantic problems such as synonymy and polysemy [2]. This research proposes to address these two challenges for building a tag-based item recommendation system by employing tensor modeling as the multi-dimensional user profile approach, and the topic model as the semantic analysis approach. The first objective is to optimize the tensor model reconstruction and to improve the model performance in generating quality rec-ommendation. A novel Tensor-based Recommendation using Probabilistic Ranking (TRPR) method [3] has been developed. Results show this method to be scalable for large datasets and outperforming the benchmarking methods in terms of accuracy. The memory efficient loop implements the n-mode block-striped (matrix) product for tensor reconstruction as an approximation of the initial tensor. The probabilistic ranking calculates the probabil-ity of users to select candidate items using their tag preference list based on the entries generated from the reconstructed tensor. The second objective is to analyse the tag semantics and utilize the outcome in building the tensor model. This research proposes to investigate the problem using topic model approach to keep the tags nature as the “social vocabulary” [4]. For the tag assignment data, topics can be generated from the occurrences of tags given for an item. However there is only limited amount of tags availa-ble to represent items as collection of topics, since an item might have only been tagged by using several tags. Consequently, the generated topics might not able to represent the items appropriately. Furthermore, given that each tag can belong to any topics with various probability scores, the occurrence of tags cannot simply be mapped by the topics to build the tensor model. A standard weighting technique will not appropriately calculate the value of tagging activity since it will define the context of an item using a tag instead of a topic.

A ‘Big Data’ Approach to Mapping the Australian Twittersphere

This chapter discusses the methodological aspects and empirical findings of a large-scale, funded project investigating public communication through social media in Australia. The project concentrates on Twitter, but we approach it as representative of broader current trends toward the integration of large datasets and computational methods into media and communication studies in general, and social media scholarship in particular. The research discussed in this chapter aims to empirically describe networks of affiliation and interest in the Australian Twittersphere, while reflecting on the methodological implications and imperatives of ‘big data’ in the humanities. Using custom network crawling technology, we have conducted a snowball crawl of Twitter accounts operated by Australian users to identify more than one million users and their follower/followee relationships, and have mapped their interconnections. In itself, the map provides an overview of the major clusters of densely interlinked users, largely centred on shared topics of interest (from politics through arts to sport) and/or sociodemographic factors (geographic origins, age groups). Our map of the Twittersphere is the first of its kind for the Australian part of the global Twitter network, and also provides a first independent and scholarly estimation of the size of the total Australian Twitter population. In combination with our investigation of participation patterns in specific thematic hashtags, the map also enables us to examine which areas of the underlying follower/followee network are activated in the discussion of specific current topics – allowing new insights into the extent to which particular topics and issues are of interest to specialised niches or to the Australian public more broadly. Specifically, we examine the Twittersphere footprint of dedicated political discussion, under the #auspol hashtag, and compare it with the heightened, broader interest in Australian politics during election campaigns, using #ausvotes; we explore the different patterns of Twitter activity across the map for major television events (the popular competitive cooking show #masterchef, the British #royalwedding, and the annual #stateoforigin Rugby League sporting contest); and we investigate the circulation of links to the articles published by a number of major Australian news organisations across the network. Such analysis, which combines the ‘big data’-informed map and a close reading of individual communicative phenomena, makes it possible to trace the dynamic formation and dissolution of issue publics against the backdrop of longer-term network connections, and the circulation of information across these follower/followee links. Such research sheds light on the communicative dynamics of Twitter as a space for mediated social interaction. Our work demonstrates the possibilities inherent in the current ‘computational turn’ (Berry, 2010) in the digital humanities, as well as adding to the development and critical examination of methodologies for dealing with ‘big data’ (boyd and Crawford, 2011). Out tools and methods for doing Twitter research, released under Creative Commons licences through our project Website, provide the basis for replicable and verifiable digital humanities research on the processes of public communication which take place through this important new social network.

Infant product-related injuries: comparing specialised injury surveillance and routine emergency department data

- Objective To explore the potential for using a basic text search of routine emergency department data to identify product-related injury in infants and to compare the patterns from routine ED data and specialised injury surveillance data. - Methods Data was sourced from the Emergency Department Information System (EDIS) and the Queensland Injury Surveillance Unit (QISU) for all injured infants between 2009 and 2011. A basic text search was developed to identify the top five infant products in QISU. Sensitivity, specificity, and positive predictive value were calculated and a refined search was used with EDIS. Results were manually reviewed to assess validity. Descriptive analysis was conducted to examine patterns between datasets. - Results The basic text search for all products showed high sensitivity and specificity, and most searches showed high positive predictive value. EDIS patterns were similar to QISU patterns with strikingly similar month-of-age injury peaks, admission proportions and types of injuries. - Conclusions This study demonstrated a capacity to identify a sample of valid cases of product-related injuries for specified products using simple text searching of routine ED data. - Implications As the capacity for large datasets grows and the capability to reliably mine text improves, opportunities for expanded sources of injury surveillance data increase. This will ultimately assist stakeholders such as consumer product safety regulators and child safety advocates to appropriately target prevention initiatives.

May the best Tweeter win : The Twitter strategies of key campaign accounts in the 2012 US election

The 2008 US election has been heralded as the first presidential election of the social media era, but took place at a time when social media were still in a state of comparative infancy; so much so that the most important platform was not Facebook or Twitter, but the purpose-built campaign site my.barackobama.com, which became the central vehicle for the most successful electoral fundraising campaign in American history. By 2012, the social media landscape had changed: Facebook and, to a somewhat lesser extent, Twitter are now well-established as the leading social media platforms in the United States, and were used extensively by the campaign organisations of both candidates. As third-party spaces controlled by independent commercial entities, however, their use necessarily differs from that of home-grown, party-controlled sites: from the point of view of the platform itself, a @BarackObama or @MittRomney is technically no different from any other account, except for the very high follower count and an exceptional volume of @mentions. In spite of the significant social media experience which Democrat and Republican campaign strategists had already accumulated during the 2008 campaign, therefore, the translation of such experience to the use of Facebook and Twitter in their 2012 incarnations still required a substantial amount of new work, experimentation, and evaluation. This chapter examines the Twitter strategies of the leading accounts operated by both campaign headquarters: the ‘personal’ candidate accounts @BarackObama and @MittRomney as well as @JoeBiden and @PaulRyanVP, and the campaign accounts @Obama2012 and @TeamRomney. Drawing on datasets which capture all tweets from and at these accounts during the final months of the campaign (from early September 2012 to the immediate aftermath of the election night), we reconstruct the campaigns’ approaches to using Twitter for electioneering from the quantitative and qualitative patterns of their activities, and explore the resonance which these accounts have found with the wider Twitter userbase. A particular focus of our investigation in this context will be on the tweeting styles of these accounts: the mixture of original messages, @replies, and retweets, and the level and nature of engagement with everyday Twitter followers. We will examine whether the accounts chose to respond (by @replying) to the messages of support or criticism which were directed at them, whether they retweeted any such messages (and whether there was any preferential retweeting of influential or – alternatively – demonstratively ordinary users), and/or whether they were used mainly to broadcast and disseminate prepared campaign messages. Our analysis will highlight any significant differences between the accounts we examine, trace changes in style over the course of the final campaign months, and correlate such stylistic differences with the respective electoral positioning of the candidates. Further, we examine the use of these accounts during moments of heightened attention (such as the presidential and vice-presidential debates, or in the context of controversies such as that caused by the publication of the Romney “47%” video; additional case studies may emerge over the remainder of the campaign) to explore how they were used to present or defend key talking points, and exploit or avert damage from campaign gaffes. A complementary analysis of the messages directed at the campaign accounts (in the form of @replies or retweets) will also provide further evidence for the extent to which these talking points were picked up and disseminated by the wider Twitter population. Finally, we also explore the use of external materials (links to articles, images, videos, and other content on the campaign sites themselves, in the mainstream media, or on other platforms) by the campaign accounts, and the resonance which these materials had with the wider follower base of these accounts. This provides an indication of the integration of Twitter into the overall campaigning process, by highlighting how the platform was used as a means of encouraging the viral spread of campaign propaganda (such as advertising materials) or of directing user attention towards favourable media coverage. By building on comprehensive, large datasets of Twitter activity (as of early October, our combined datasets comprise some 3.8 million tweets) which we process and analyse using custom-designed social media analytics tools, and by using our initial quantitative analysis to guide further qualitative evaluation of Twitter activity around these campaign accounts, we are able to provide an in-depth picture of the use of Twitter in political campaigning during the 2012 US election which will provide detailed new insights social media use in contemporary elections. This analysis will then also be able to serve as a touchstone for the analysis of social media use in subsequent elections, in the USA as well as in other developed nations where Twitter and other social media platforms are utilised in electioneering.