In early 2020, we substantially updated the underlying ATNIX architecture. Most importantly, we now reduce the various URLs that may lead to the same story to a single canonical URL, in order to arrive at a reliable count for how often a story has been shared (rather than just how often a particular URL variation has been shared). This has become necessary because many news sites incorporate part of the story headline into its URL – but headlines may change after publication, and so multiple different URLs may point to the same story in the end.

ATNIX tracks the sharing of stories from most major Australian news outlets – from ABC News to New Matilda and beyond. We exclude international outlets with an Australian presence (such as The Guardian Australia or Mail Online Australia), because the majority of their content originates from outside of Australia, but we continue to include The Conversation because it remains Australian-based and sources a substantial amount of its content from Australian authors. The data gathered for ATNIX include all tweets, by Australian as well as international Twitter accounts, that link to the domains of these Australian news outlets. From these, we exclude links to their homepages as well as to non-news content.

The ATNIX Twitter account (@_ATNIX_) posts half-daily, daily, and weekly updates on trending Australian news stories, and ATNIX also provides an interactive dashboard with live and historical data on sharing patterns for Australian news, at and above. In earlier years, ATNIX analysis was published in a regular column in The Conversation (https://theconversation.com/columns/axel-bruns-1433).

The Australian Twitter News Index has reported on patterns in the sharing of Australian news content through Twitter for many years; it has documented the overall stability of Australian Twitter users’ preferences for specific news sources (especially ABC News and the Sydney Morning Herald), tested the overlap in content-sharing audiences between news outlets with different editorial and ideological orientations, and reported on the most popular news stories during specific timeframes.

Datasets analogous to ATNIX are also being collected for Germany, the Nordic countries, Spain, and a selection of suspected sources of mis- and disinformation.

Key scholarly discussions of ATNIX and its data can be found in:

Bruns, A. (2016). Big Data Analysis. In T. Witschge, C. W. Anderson, D. Domingo, & A. Hermida (Eds.), The Sage Handbook of Digital Journalism (pp. 509-527). Sage. https://eprints.qut.edu.au/102642/

Bruns, A. (2017). Making Audience Engagement Visible: Publics for Journalism on Social Media Platforms. In B. Franklin & S. A. Eldridge II (Eds.), The Routledge Companion to Digital Journalism Studies (pp. 325-334). Routledge. https://eprints.qut.edu.au/102644/

For further questions about ATNIX and its datasets, please contact the project leader, Prof. Axel Bruns.

The Nordicom Review article was published under an open access licence – here’s the full abstract:

Previous research into social media platforms has often focused on the exceptional: key moments in politics, sports or crisis communication. For Twitter, it has usually centred on hashtags or keywords. Routine and everyday social media practices remain underexamined as a result; the literature has overrepresented the loudest voices: those users who contribute actively to popular hashtags. This article addresses this imbalance by exploring in depth the day-to-day patterns of activity within the Australian Twittersphere for a 24-hour period in March 2017. We focus especially on the previously less visible everyday social media practices that this shift in perspective reveals. This provides critical new insights into where, and how, to look for evidence of onlife traces in a systematic way.

This builds on our prior work to identify Australian Twitter accounts and map the network structure of the Australian Twittersphere (covered at The Conversation and published in Social Media + Society), and tracks the public posting activities of some 500,000 Australian Twitter accounts on a continuous, real-time basis. For this general overview, we’re pulling out the major hashtags and the most mentioned accounts (counting both @mentions and retweets) – but of course the underlying dataset captures far more than this.

If you’re interested in further research that builds on this dataset, please get in touch!

(Click ‘full screen’ to enlarge.)

The Australian Twittersphere in 2016: Mapping the Follower/Followee Network

Twitter is now a key platform for public communication between a diverse range of participants, but the overall shape of the communication network it provides remains largely unknown. This article provides a detailed overview of the network structure of the Australian Twittersphere and identifies the thematic drivers of the key clusters within the network. We identify some 3.72 million Australian Twitter accounts and map the follower/followee connections between the 255,000 most connected accounts; we utilize community detection algorithms to identify the major clusters within this network and examine their account populations to identify their constitutive themes; we examine account creation dates and reconstruct a timeline for the Twitter adoption process among different communities; and we examine lifetime and recent tweeting patterns to determine the historically and currently most active clusters in the network. In combination, this offers the first rigorous and comprehensive study of the network structure of an entire national Twittersphere.

A summary of some of the study’s key findings was published in The Conversation in May 2017. Meanwhile, a paper by Axel Bruns at the Future of Journalism conference in Cardiff in September 2017 built on this new Twittersphere map to test for the existence of echo chambers and filter bubbles in Australian Twitter – and found little evidence to support the thesis:

Echo Chamber? What Echo Chamber? Reviewing the Evidence

The success of political movements that appear to be immune to any factual evidence that contradicts their claims – from the Brexiteers to the ‘alt-right’, neo-fascist groups supporting Donald Trump – has reinvigorated claims that social media spaces constitute so-called ‘filter bubbles’ or ‘echo chambers’. But while such claims may appear intuitively true to politicians and journalists – who have themselves been accused of living in filter bubbles –, the evidence that ordinary users experience their everyday social media environments as echo chambers is far more limited.

For instance, a 2016 Pew Center study has shown that only 23% of U.S. users on Facebook and 17% on Twitter now say with confidence that most of their contacts’ views are similar to their own. 20% have changed their minds about a political or social issue because of interactions on social media. Similarly, large-scale studies of follower and interaction networks on Twitter show that national Twitterspheres are often thoroughly interconnected and facilitate the flow of information across boundaries of personal ideology and interest, except for a few especially hardcore partisan communities.

Building on new, comprehensive data from a project that maps and tracks interactions between 4 million accounts in the Australian Twittersphere, this paper explores in detail the evidence for the existence of echo chambers in that country. It thereby moves the present debate beyond a merely anecdotal footing, and offers a more reliable assessment of the ‘echo chamber’ threat.

Stephen Harrington, Axel Bruns, and Tim Highfield. “Infotainment and the Impact of ‘Connective Action’: The Case of #MilkedDry.” Paper presented at the Australia New Zealand Communication Association conference, Sydney, 6 July 2017.

Axel Bruns, Brenda Moon, and Ehsan Dehghan. “Dynamics of a Scandal: The Centrelink Robodebt Affair on Twitter.” Paper presented at the Australia New Zealand Communication Association conference, Sydney, 7 July 2017.

Part of the problem here is that we all design our own filter bubbles, as the journalism researcher Paul Bradshaw has put it: what two random users see of Twitter might be entirely different, depending on what other accounts they choose to follow. If all you ever see is food porn, perhaps you’d care to make some new connections. Or perhaps that’s what you’re there for.

But if we could look beyond our own, personal networks, what would we see? What are the major drivers of Twitter take-up, in Australia and elsewhere? Do we connect around shared interests, shared location, or pre-existing offline relationships? And when, in the eleven-year history of the platform, did these structures form?

These are the questions that guided a new, long-term study of the Australian national Twittersphere that my colleagues and I have undertaken. Drawing on TrISMA, a major multi-institutional facility for social media analytics, we identified some 3.7 million Australian Twitter accounts in existence by early 2016, and captured the 167 million follower/followee connections they have amongst each other.

Twitter took off in Australia in 2009, some three years after its launch, and saw a fairly steady sign-up rate of 1,000-2,000 new accounts between 2010 and 2014. Growth has slowed since then, and this may indicate market saturation. There are a number of obvious spikes in new account sign-ups, too: the series of natural disasters in early 2011 attracts users to the platform who recognise its role in crisis communication, and the political turmoil of 2013 also seems to drive take-up.

A major spike in 2015 appears to coincide with the devastating Nepal earthquake, but we’ve yet to determine why that event would lead to new Twitter accounts being created in Australia.

To focus in on the core parts of the network, we further filtered this to accounts that have at least 1,000 connections in the global Twittersphere, which left us with the 255,000 best-connected accounts. We visualised their network using Gephi’s Force Atlas 2 algorithm, which places accounts close to each other if they share many connections, and further apart if they are only poorly connected.

The network map shows clear clustering tendencies: dense regions, where many accounts are closely connected, are separated from each other by lower-density spaces. We systematically examined these clusters, and labelled them based on the overarching themes that emerged from an analysis of the account profiles in each cluster. The result is a kind of birds-eye view of the Twitter landscape, from politics to popular culture and from education to sports.

Perhaps surprisingly, accounts connecting around teen culture make up the largest part of this network: 61,000 of our 255,000 accounts are located here. Other major clusters include aspirational accounts (these include self-declared social media gurus, self-improvement and life-coaching practitioners, and others who sought to use Twitter for professional betterment), at 26,000 accounts; sports, with 25,000 accounts (including distinct sub-clusters for cycling and horse racing); and netizens, technologists, and software developers (17,000 accounts).

Shared interests emerge from this as the central drivers of our connections on Twitter: for the most part, we follow others because of the topics they cover, not because they’re from the same city or state or because we already know them offline. An equivalent map for Facebook, where connections are much more strongly based on prior acquaintance, would likely look very different.

We further found that these accounts also arrived on Twitter at very different times: both netizen and aspirational accounts were created very early in the history of the platform. As expected, netizens constituted the vast majority of Australia’s early adopters, with aspirational accounts close behind; fully half of the population in both these clusters had arrived on Twitter by mid-2010. Sports took a year longer, and may well have been helped along by Twitter Australia itself as it reached out to key sporting codes to get their teams and players signed up.

By contrast, the teen culture accounts arrived a great deal later. It took until mid-2012 until half that cluster’s population had joined – the teen invasion of Twitter represents a secondary adoption event, following the first big influx of Australian users in 2009/10. Here, too, we suspect active encouragement from key bands like One Direction and Five Seconds of Summer as a major driver.

In spite of Twitter’s reputation as a space for political debate and agitation, politics attracts only some 13,000 accounts (including 1,500 that form a separate, staunchly right-wing cluster); there’s a great deal more to Twitter than political argument.

But if all you ever see on Twitter is partisan bickering, there may be a reason: per capita, the political accounts are some of the most active in the Australian Twittersphere. Over their lifetimes, they’ve posted an average of 7.2 tweets per day (and the accounts in the hard right cluster even manage 12.5 per day); in the turbulent first quarter of 2017, those averages are even higher. Most of the other major cluster communities have managed less than half that work rate; historically, only the teen culture accounts have been similarly active.

Twitter is what its users make it, and Australian users have made it a diverse and dynamic place, even if perhaps they’re less aware of each other than they should be. As users, we should step beyond our networks more often, to avoid becoming trapped in our own filter bubbles – and this goes doubly for politicians, journalists, and others who now treat their immediate Twitter networks as an instant source of popular opinion.

And as a company, Twitter too has much work to do to enable its users to experience the full variety of networked communication and culture that the platform has to offer. Changes to how it recommends new accounts to follow, and how it reveals trending topics outside of our existing networks, could help a great deal in combatting the threat of getting stuck in your own filter bubble.

It doesn’t stop there, of course. We can only speculate what the equivalent networks for Facebook, Instagram, or Snapchat would look like, and what they might tell us about how people are using these platforms.

(An edited version of this article was published in The Conversation on 3 May 2017.)

I’m now getting ready to get ATNIX started up again, though, and hopefully to make some further additions that will prove useful in the longer term. To get us started, I thought it might be useful to post a long-term overview of ATNIX trends since we started the index in mid-2012. Over the past two years, we’ve seen a growing adoption of Twitter in Australia, to a point where there are now more than 2.8 million accounts in the Australian Twittersphere – and it seems logical that this would also manifest in changes to the sharing patterns for Australian news sites on Twitter.

Indeed, the total volume of tweets sharing links to Australian news sites has increased during these two years – as has, it should be noted, the number of news sites we’ve tracked. In total, since mid-2012 (and allowing for a handful of server outages), we’ve captured some 20 million tweets in total, containing more than 24.5 million URLs. And those numbers have increased steadily: while in July 2012, we saw a total of 677,000 tweets linking to our Australian news sites, by July 2014 that number had grown to more than one million. (In fact, 2014 has seen particularly strong growth, perhaps due to the substantial confluence of various domestic and international events and crises.)

Broken down across the 35 Australian news and opinion sites we are currently tracking, these patterns look as follows (click to enlarge, and ignore the obvious drop-outs due to server maintenance in November 2013):

For long-term followers of our ATNIX data, it is immediately evident that the overall rankings amongst the major news sites have remained largely stable: ABC News and the Sydney Morning Herald remain the most widely shared news sites in Australia by some margin (and, it seems, by a margin that continues to increase relative to their nearest competitors). In the second tier, The Age and news.com.au are similarly running neck-and-neck. And they are followed, finally, by the rest of the field, with some of those sites occasionally recording major spikes due to the viral dissemination of single stories.

A closer look reveals a few more interesting patterns, however. The SMH appears to have recovered from a lengthy slump in popularity that began in early 2013, which saw it fall back from ABC News’ tail, and since April 2014 has been shadowing its major competitor much more closely once again. Amongst the opinion and commentary sites, The Conversation is the obvious market leader, though this is also boosted by its new-found transnational reach, with strong take-up in the UK and elsewhere – and it should be noted that following the site’s conversion from a .edu.au to a .com address we missed some months of data early this year, so its lead over nearest competitor Crikey would likely be even greater. And overall, the greatest spike in news sharing activity occurred, unsurprisingly, during the last federal election, when we captured more than 50,000 tweets linking to ABC News for the election week alone.

Sadly absent from this chart, however, are Guardian Australia and Daily Mail Australia. Due to their lack of a dedicated Australian domain, or of any other markers identifying their Australian coverage, we’re unable to separate Australia-specific news sharing activities from the global Guardian and Mail brands, and therefore cannot include them here. (We’re choosing to include The Conversation despite its now international audience, however, because it originated and continues to be substantively based in Australia.) Eventually, as we develop our data gathering approach further, we hope to develop the methods to better identify Australian-based sharing of news from these sources.

Introducing Experian Hitwise Data

As we develop ATNIX further, we also hope to place it into a wider context by comparing these Twitter-based news sharing patterns with reading and sharing activities elsewhere. We’ll soon attempt to tackle Facebook, but for now, here’s a glimpse of a very different data source: Experian Hitwise. Experian Marketing Services collects anonymous data at ISP level through opt-in panels about the Web searching and browsing patterns of Australian Internet users, and in the graph below I’ve compiled the site visit statistics for the same sites which we are tracking as part of ATNIX, for the same timeframe:

Total visits to Australian news and opinion sites, July 2012 to September 2014. Data courtesy of Experian Marketing Services Australia.

Once again, a significant rise in the total number of visits to news sites by Australian Internet users since the start of 2014 is evident, corresponding to a similar rise in news sharing during this time; we’re also seeing a matching dip in late April/early May, during the Easter / ANZAC Day holiday period. However, the ranking of news sites is markedly different: since early 2014, the market leader in Australian online news is news.com.au, even if such leadership doesn’t result in a similarly strong result in news sharing as we measure it through ATNIX. Conversely, ATNIX leader ABC News ranks ‘only’ fifth amongst the most read news sites in Australia.

Amongst the opinion and commentary sites, The Conversation and Crikey lead the Experian Hitwise rankings, too, but the rest of the leaderboard is structured quite differently. This is probably an indication of the respective positioning of these sites: to attract a loyal readership in their own right, to encourage the viral distribution of their articles, or both. Experian Hitwise records a surprisingly strong readership for The Morning Bulletin, for example, while ATNIX does not show its content to be very widely shared through Twitter; conversely, New Matilda content is widely shared, but according to the Experian Hitwise figures it does not seem to have a very large regular audience.

And finally, the Experian Hitwise numbers also provide us with a glimpse of Guardian Australia’s and Daily Mail Australia’s market positioning: by late September they’ve managed to rise to eight and fifth place on the Experian Hitwise chart, respectively, and continue to trend gradually upwards. We’ll watch their further development with interest.

Standard background information: ATNIX is based on tracking all tweets which contain links pointing to the URLs of a large selection of leading Australian news and opinion sites (even if those links have been shortened at some point). Datasets for those sites which cover more than just news and opinion (abc.net.au, sbs.com.au, ninemsn.com.au) are filtered to exclude the non-news sections of those sites (e.g. abc.net.au/tv, catchup.ninemsn.com.au). Data on Australian Internet users’ news browsing patterns are provided courtesy of Experian Marketing Services Australia. This research is supported by the ARC Future Fellowship project “Understanding Intermedia Information Flows in the Australian Online Public Sphere”.

Given this popularity, and a significant overlap between the target market for Big Brother viewers and the social media platforms, it will be interesting to observe the extent to which social media activity (and perhaps, eventually, sentiment) acts as a predictor for votes on the show. In this blog, following the first round of nominations, first eviction and the first round of single nominations, we are going to look to the data from the last 2.5 weeks to try to test whether social media activity acts as a predictor of public votes.

So far, at least, it has been a mixed bag, but let’s start with the positive; the public vote for the ‘Perfect Pair’ dance competition, in which the winners were awarded $30,000, was held between the final two pairs – Lawson and Aisha & Dion and Jason. The public then voted for the pair with the best dance through JumpIn, but did they actually just vote for their favourite pair? If we use social media activity as a barometer, it seems that could be the case. Our data showed a tight race, which Lawson & Aisha just pipped, and indeed the public vote came back 51.8% in favour of Lawson & Aisha. Perhaps, if they had been up against, say, Travis and Cat – who were hardly mentioned this week – they would have won by even more:

Lawson also tells an interesting story in the overall polling; as seen in the chart below which highlights the running total for all housemates; largely anonymous until the dance-off and his decision to give Aisha the lions share of the prize money ($20,000) was rewarded in the social media volume.

Below is a running total of Twitter mentions for the pairs since launch night, however we will focus on the last week’s long-winded and highly debated eviction process for the time being. Nominees made up 5 of the six most talked about housemates on the night before the eviction process began, and the ones not being talked about were being carried by their partner based on the pairs table:

We can of course ask some other interesting questions from these charts: where were Skye and Lisa when they were ‘saved’? Were Jake and Gemma losers in the public vote due to anonymity, or hatred? What caused David and Sandra to be saved, when they were virtually anonymous through the first week, and only talked about subsequently in regard to David’s chauvinistic comments. Was it better for David to be hated, rather than not talked about at all? Related to this, there is the question of screen time and popularity inside the house, allowing us to address what went wrong for Gemma this week, given her achieved intent to secure airtime?

Up for eviction this week were Skye & Lisa, Jake & Gemma, Travis & Cat and David & Sandra. Ever since the Katie & Priya first week fiasco, Skye & Lisa have been by far the most talked about pair of the season and consequently were saved on Monday night as per our prediction based on the previous graph, with Skye & Lisa the most popular pair on the 22nd September. Interesting here, however, is that Gemma & Jake were the pair with the second most social media activity, and the most popular during the nomination period, indicating that the sentiment will also be a significant factor in creating further predictions.

While we have our own tool monitoring Big Brother discussion (http://bigbrother.thehypometer.com), Channel 9 (Mi9/JumpIn) have also launched a counter, the “Big Brother Radar”, which captures tweets and Facebook statuses by those who seek, deliberately, to be noticed by the radar using official C9 hashtags (e.g. #BBAUGemma). Our tool, by contrast, attempts to measure the underlying volume of discussion (and, by possible inference, interest) in the competitors as a whole, on social media.

Going forward, we hypothesise that those housemates who the public have no interest in will be those who struggle in a ‘vote to save’ format. That said, it’s probably not advisable to bet based on this information. It may be that the Radar format serves as a better prediction of those likely to be evicted (i.e. the effort to post with the correct hashtag is correlated to the effort to vote), it may be that sentiment proves highly significant, or indeed it may be that social media is not a good barometer of the BB voting public. Whichever of these proves to be the case however, the data is sure to be interesting.

Finally, it is worth noting that one of the problems of a lack of live feed – which we have ranted about previously – and indeed this year any live updates at all is that it allows producers to largely control the message; hence, social media reaction largely follows the amount of airtime given to contestants and the plot lines developed, much like a soap. By contrast in the USA, with 4 live camera views running 24 hours a day, users are able to create and share their own storylines about the housemates — generating ‘hype’ for the show which we do not see here. In Australian Big Brother we are told what to think, and we’ll leave it as an exercise for the reader how that reflects on wider society. Finally, we’ll leave you with a running total of the housemates mentions to date, where Skye continues to lead the way:

Methods

First things first: here’s our methodology for getting to this point. Over the course of several months in 2013, the tools developed by our data scientist Troy Sadkowsky used the Twitter API to access the publicly available profile information for each account then in existence; we simply pinged every user ID from 0 through to (at that point) upwards of 2 billion, and recorded the information returned. This resulted in data for some 750 million accounts – the size of the global Twitter userbase (or more precisely, account base) around September 2013. (We’ll share some analysis of the global trends in Twitter account sign-ups in a separate post in the near future.)

This comprehensive snapshot of global Twitter accounts provides us with an opportunity to go looking specifically for Australian users. To do so, we drew on three key elements of each user profile: the free-text profile description and location fields as entered by the account creator, as well as the profile timezone they chose from the pull-down menu of presets offered by Twitter. On the basis of the latter, we selected all users who had chosen one of the eight state-based Australian timezone options, while for the former two fields, we developed a long list of search terms relating to Australian towns, cities, and states, and to Australia itself, using a number of common variations. Any account that matched our criteria for “Australianness” in any of these three fields has been included in our selection.

To go through the full list of search terms would take up another post, but we worked with a list of the 50-odd largest cities in Australia, added in a handful of popular variations, included the state names and their abbreviations, and also used terms such as “Australia”, “Stralya”, “down under”, and others. Following a test run, we further refined these terms, to include popular misspellings (“Austalia”, “Tasmainia”) and remove false positives. This turned out to be a somewhat time-consuming exercise: many place names in Australia are re-used from Europe (“Perth”, “Ipswich”) or duplicated in other new world countries (Brisbane, California; Victoria, British Columbia); some Australian place names also appear in popular media (some users claim to be from the “City of Townsville” or indeed the “Ciudad de Townsville” in homage to the Powerpuff Girls, or from Finding Nemo’s “42 Wallaby Way, Sydney”). Where possible we’ve filtered out any false positives which could be clearly identified.

In the end, this process of filtering the total dataset of over 750 million Twitter accounts left us with some 2.8 million accounts whom we are confident to classify as ‘Australian’ for the purposes of this study. For many of these, we are also able to assign a likely state and/or city, based on which of our search terms helped identify the account; here, we give greatest credence to the information contained in the location field of the Twitter profile, followed by description and timezone. Where we identified users only based on their timezone, we have assigned a state, but have refrained from assigning them to the state’s capital city.

Inevitably, some false positives will remain in our dataset, and some accounts will be miscategorised – “Sydneysider now living in Melbourne” or “Australian in New York” may lead to false location assignments, and descriptions like “Korean student in Brisbane” or even “Dreaming of travelling through Australia” would have matched our search terms, but do not relate to the accounts of Australian users in a narrow sense. However, given the size of the total dataset our best-match approach using automated processes is the best option available to us, and I’d guess that some 90-95% of the accounts we’ve matched are genuine Australian users: either Australians in Australia, Australians elsewhere in the world, or non-Australians living in Australia. The outliers from this population are likely to show up in our further analysis, too.

There will also be some false negatives, of course: accounts which give no indication of their Australian connections anywhere in their location, description, or timezone details (including users who have filled in none of their profile details at all). It seems likely that the greatest number of these will be amongst the most recently registered accounts (whose owners may not yet have had a chance to fully customise their Twitter settings), so we’ll largely ignore this group for now – we’ll re-run our survey of the total Twitter userbase at some point in the future to examine how these accounts may have developed, as well as to gather data on the accounts which were created after our initial data-gathering exercise finished in September 2013.

Findings

By the end of August 2013, then, the Australian Twittersphere included some 2.79 million accounts, by our criteria. Per capita, using the Australian Bureau of Statistics’ figures for September 2013, this would translate to a 12% sign-up rate, though that figure must be viewed with some caution: some Twitter users will operate multiple accounts (e.g. for private and professional use), while in other cases several users will share the same group account. This is why we’re careful here to speak of 2.79 million accounts, rather than users.

This figure is in line with existing reports and guesstimates for the size of the Australian Twittersphere, if somewhat below the 4 million Australian accounts that Twitter, Inc. itself apparently boasted some months ago. Figures from the company itself should always be taken with a grain of salt, of course; they’re largely released for corporate promotion reasons, and may well reflect the total number of Australian-based accounts ever created, rather than the number of accounts which are still in existence at present (which is what we measured). On the other hand, there is also an unknown number of accounts which our methods would not identify as Australian, based on publicly available profile details, but which Twitter, Inc. (which would have identified the IP address from which a Twitter account was created) would classify as Australian. This also explains some of the discrepancy in numbers.

Here’s how that population has grown month by month over the seven years covered by our dataset (click on the images for larger versions):

From a slow start over the first couple of years (which is similar outside of Australia, too), there’s finally a sudden and rapid rise in new registrations per month in early 2009, peaking at over 100,000 new account registrations each in March and April 2009. (And there may well have been more than this: the 100,000+ accounts we see joining in these months are only those which were still in existence when we gathered our data in late 2013, of course.) From this early excitement, things slow down considerably towards the end of 2009 – and then trends start to point upwards again: the average number of new accounts joining per month during the following years is somewhere around 40-50,000. Finally, there is a substantial increase in new registrations in August 2013; this may be partly related to the impending federal election, but probably also reflects the fact that Twitter’s spam bot-checking systems may not yet have had a chance to remove any offending new accounts.

We should also note, though, that what our data cannot (yet) tell us is the number of accounts which are being deleted each month, and how those deletions compare to the influx of new accounts. We’ll have a better indication of this after the next iteration of our survey, which will allow us to examine the discrepancies between the two datasets: accounts present in the September 2013 dataset but absent from the new iteration must have been deleted (by their owners, or by Twitter, Inc.) in the meantime.

State-by-state patterns vary quite considerably at times. There are unusual spikes in ACT and Queensland account registrations between April and September 2012, for example which do not appear to be motivated by specific local events; ACT sign-ups per month rise from below 1,000 to over 4,000 accounts during that period, for example. From a preliminary review of the accounts which joined during that time, it appears that a considerable number of them belong to fans of The Janoskians, One Direction, and other teen bands, so perhaps there was a concerted effort by some of these bands to get their fans on Twitter?

Other spikes are clearly driven by more sinister motives. The large spike in generically ‘Australian’ accounts in January 2013 is caused almost entirely by a large number of spam bots being created at virtually the same time, for example: of the 1,106 new accounts on 16 January 2013 alone, we counted 170 accounts claiming to be “Australia’s support member for the Global Information Network”; 153 offering “Australian Business for Sale listings”; 155 promoting “software and services in Singapore. Australia. China and Japan”; and 164 accounts claiming to be an “Independent Mortgage Broker in Australia” – that’s almost two thirds of the ‘Australian’ accounts for that day. Clearly Twitter’s spam account filters still have some way to go.

But genuine events in the world also result in increased sign-ups. During the first quarter of 2011, for example, we see a considerable spike in new Queensland-based accounts on 11 and 12 January, as floodwaters threaten inner-city Brisbane, and during the following days; in Victoria, New South Wales, and other states the sign-up rate also increases notably. Similarly, as a devastating earthquake hits Christchurch, New Zealand, on 22 February, Australians also sign up in larger numbers than usual. The pattern does not repeat (other than perhaps in Queensland, once again) following the 11 March earthquake and tsunami on the east coast of Japan, however.

The graph above also shows a considerable dip in new registrations on 18 February 2011 – this may well be due to an outage in Twitter’s account registration systems.

The geographical distribution of these accounts should necessarily be treated with a certain degree of caution, given the vagaries of correctly identifying cities and states from the free text provided by users in the location and description fields. However, the patterns we’re able to determine from our best guess at the likely location of each user do reflect both the overall distribution of the Australian population and the relative likelihood (based on infrastructural and socioeconomic factors) of local residents joining Twitter that we would expect to see:

The major population centres are clearly leading the way. Sign-up rates per capita seem to be strongest in the state capitals and on the Gold Coast, but this may be an artefact of our approach, which focussed on identifying mentions of the 50-odd major population centres in Australia in the location and description fields of users’ Twitter profiles. Because of the greater national and international recognition of such centres, city users may state that they’re from state capitals while those from small rural and regional locations might just mention their state. In a further iteration of our work, we’ll check against a longer list of localities in Australia, and the patterns may well change.

We’re on more solid ground when we examine the sign-up rates for each state. This aggregates users who name specific cities with those who only specify a state, and accounts for some 2.4 million of our total 2.8 million identified accounts – about 420,000 accounts we identified as ‘Australian’ referenced only generic terms (“Australia”, “down under”, etc.), but did not include any more specific location details.

For most states, the sign-up rate ranges between 8 and 11 per cent, with Queensland and (perhaps somewhat surprisingly) the Northern Territory taking the lead of this group. There are likely to be any number of factors which have resulted in these slight differences in Twitter adoption across the country; for Queensland, for example, the well-publicised utility of Twitter during recent natural disasters may well have contributed to an above-average take-up. If the 420,000 accounts which we could not allocate to any specific state were distributed proportional to the states’ population figures, this would boost each sign-up rate by another 1.8 percentage points, incidentally.

But the major story here is of course the ACT, which records a whopping per capita take-up rate of 30%. We’ll have to look more closely into what factors are responsible for this pattern – but so far we have not seen any indications that an unusually large number of false positives have slipped through our net. There are, however, unusually many accounts whose only identifying feature is their ACT timezone setting, and it is always possible that people from other UTC+10 timezones (for example in the northern hemisphere) might have chosen the ACT timezone rather than searching for their own options in the pull-down menu available on the Twitter site.

Another factor that might drive the abnormally high number of accounts with some relation to the ACT is a combination of the socioeconomic make-up of the ACT population, and the fact that (as the seat of the federal government) there will be a very substantial number of organisational accounts, politicians, journalists, public servants, and other likely Twitter adopters in Canberra and surrounds. Additionally, there may also be a significant discrepancy between the number of formally registered ACT residents and the number of people who actually live and/or work in Canberra at least part of the time.

If we break down state numbers per city, the capital cities unsurprisingly account for the majority of Twitter accounts. There are also many accounts for which a city couldn’t be determined – these are accounts which merely chose an Australian timezone, which named only their state in the location or description field, or which stated a location other than the 50-odd most populous Australian cities we searched for. Further, though, it is also notable that Queensland’s Twitter population appears to be most geographically dispersed: in addition to the Gold Coast (which is a major population centre in its own right, of course), it also boasts the widest range of other centres with Twitter userbases numbering above 1,000 accounts. This is largely reflecting the population distribution across various regional centres in central and far north Queensland, but may also point to the useful role Twitter now regularly plays during Queensland’s summer storm season.

So much for a first overview of the overall figures. Over the next months, we’ll delve much more deeply into the patterns which this massive dataset of Australian Twitter accounts reveals – and we’ll also develop a number of approaches to mapping the follower/followee networks of this Twitter population.

(Cross-posted from Mapping Online Publics.)

A Quick Look Back

So, as we left things, we put out a few predictions for the week beginning January 20, as follows:

So, some hits and misses here.. The most concerning of these are Pretty Little Liars & Ravenswood, so let’s have a closer look at those, starting with Ravenswood. Essentially, I think this was an over-prediction influenced by the second half premiere, which rated at 181,200 tweets. We have now added an adjustment for these type of shows, that will highlight second half premieres in our predictor, with the option to manually exclude them where they rate significantly different. More concerning was “Pretty Little Liars”, which wasn’t a second half premiere but a regular show. Here, I think we had the reverse problem. Because there weren’t enough shows in the past two months (only 1 after the second half premiere), our prediction algorithm defaulted to include episodes from the previous series, where 300-400k tweets was the norm. Combining this with the 488,000 tweets from last week, this seemed a reasonable estimate, but the show actually nearly doubled that performance — it will be interesting to see how Pretty Little Liars fairs over the coming weeks.

The other errors of over 10% were both reality shows, and here I’m tempted to put the predictions on hold until we’ve established a measure of season context. Just as Big Brother follows a weekly cycle of Daily Shows, Nomination Shows and Eviction Shows, shows like American Idol, The Voice etc follow a somewhat standard season format between auditions, performances, eliminations and so forth. Modelling this is on my to-do list, but it’s after the establishment of the Twitter Excitement Index and indeed the modelling of regular season shows with their premieres, finales and other formats. Time and other priorities didn’t allow us to get predictions out for this week, and I won’t do so retrospectively, but we’ll get some out either late this week or early next.

Twitter Excitement Index

One thing we did finish last week was the establishment of our Twitter Excitement Index, which is a measure of the volatility of conversation on Twitter based on the principles of Brian Burke’s Excitement Index at Advanced NFL Stats. Over a few weeks Katie Prowd and I went through a few variations on this approach, looking at different measures to see which best captured the patterns in Twitter conversation, and I also owe thanks to Patrik Wikstrom for his help in tweaking our statistical approach. Essentially though, the theory here is that if you have two shows, they may both average 100 tweets per minute, but have very different engagement levels, as seen in this dummy data:

In the first graph, the activity is ‘spiky’, that is, every other minute people are prompted to tweet, while in the second there is an underlying level of 100 tweets, with minimal variation around that average, suggesting a constant stream of conversation but no particular moments which provoke users to tweet. These should be at opposite ends of the spectrum, and with our Twitter Excitement Index, the top graph would see a TEI of 9.9, while the second gets a TEI of 0.5. The scale here has been calculated to vary between 0 and 10 for presentation purposes, although in practice from our test data it would seem very unusual for a show to achieve a TEI of over 5.

The Australian Premieres

We are currently working on adjusting our metrics to work with Australian television, and this weeks premieres of The Block and My Kitchen Rules gave a good first test for this approach. But first, let’s cover the basics, and look at how these shows performed on Twitter:

As you can see here, My Kitchen Rules clearly won the night, both in terms of total tweets and audience peaks, from The Block in second place. The third line here represents The Biggest Loser, which wasn’t premiering but did air an episode in competition with the other two reality shows. The twitter audience for The Biggest Loser has fallen off a bit since its premiere, but this performance must be considered low by any standards. A similar picture was seen in the TV ratings, with My Kitchen Rules reaching 2.4m viewers, The Block 1.1m, and The Biggest Loser just 560k. However, it’s interesting to compare these to other recent reality ‘events’:

What is evident here is that Big Brother still owns the crown for Australian reality television, at least on Twitter, with both the Premiere and Finale easily outperforming the launch episode of My Kitchen Rules. For Channel 9, the performance of The Block suggests that the reality show success they achieved with Big Brother is not easily transferable to other shows, and for Channel 10 both the performance of Masterchef and Biggest Loser must be a cause of some concern… A similar pattern can be seen in the unique audience for Monday’s shows (n.b. the percentages add up to over 100% because of viewers who tweeted about multiple shows):

Finally, the launch of these shows also gave us an opportunity to test some of our new metrics, and after a slight hiccup caused by time zone confusion (the TEI saw the drop-off to QLD discussion as a good thing for the show, viewing NSW levels as large spikes), we ended up giving a slight win to MKR, although both ranked below the season launch of The Biggest Loser. The TEI does not take into account tweet volume, only measuring the volatility of conversation among the audience that *did* watch the show, and so should generally be read in combination with the audience size to properly understand the conversation around a show.

Mumbrella also covered some of this, and you can view their article here.

And so wraps another week in Telemetrics.. until next time!