Analysis Crisis Twitter — Snurb, 17 January 2011
The Queensland Floods on Twitter: A Brief First Look

Update: added another graph showing the total number of tweets from leading accounts.

Queensland has just experienced a once-in-decades flood event, with the capital Brisbane (where we’re based) hit especially hard. Social media like Twitter and Facebook played an important role in getting information out and organising rescue, relief, and recovery operations.

We’ll have a much closer look at the role of these platforms during the height of the crisis at a later stage, when we find the time – for now, I wanted to post a quick overview of the level of Twitter activity at least. This graph shows tweets using the #qldflood(s) hashtags between 11 and 14 January 2011 (retrieved via Twapperkeeper):

image

(11 Jan. 2011 was the day that flash floods washed through Toowoomba and devastated towns in the Lockyer Valley. In the early hours of 12 Jan. 2011 they reached Brisbane, peaking early on 13 January.)

Also worth noting: the most prominent participants in #qldflood(s). Here are the accounts which received the most @replies (including old-style retweets – click on the graphs for full size, as usual):

image 

Somewhat surprisingly (since it was relatively unknown before the crisis), the Queensland Police Service’s @QPSmedia account emerges as a clear frontrunner, with media outlets @abcnews (the national public broadcaster) and @couriermail (the local Brisbane newspaper) also very prominent. Key news sources remain important even in social media, in other words, especially during times of crisis.

By comparison, fourth-placed @vonbunnie appears here only because her tweet

RT @vonbunnie: For every RT of this tweet I will donate 10 cents towards the Aussie Queensland flood appeal. Help me out. #QLDFloods

was, unsurprisingly, widely retweeted (and I have not confirmed whether the 372 retweets during these four days did result in a $37.20 donation, or whether @lilithia made good on her promise of $1 per retweet, totalling $291 over the same period).

Also interesting to see celebrities such as @Alyssa_Milano and @Pink featured prominently. Again, this is mainly through retweets, for example of

RT @Pink: My heart and prayers go out to all my Aussies I’m praying for your safety and health and hearts. #qldflood

- the effects of star power! (And note that this only includes old-style retweets, i.e. ‘RT @Pink …’ – not new-style retweets via the ‘retweet’ button.)

But here, we’re already starting to enter a very long tail – almost 400 users in the #qldflood(s) hashtag community received ten or more @replies over these four days, so there was plenty of discussion and retweeting going on across the board.

Update:

From Jonathan Este at Walkley Magazine comes the request for a graph of the total number of tweets made by the most active contributors to #qldflood(s) – which is a really great idea, and easy to do. Here it is, again for the period of 11 to 14 Jan. 2011:

image

Some very active automated feeds there that retweeted a large number of #qldfloods posts (@thebigwetfeed, @qldfloodfeed), but also plenty of regular users pitching in to make sure important messages get widespread distribution. And clearly, emergency services and media organisations are also doing their bit – @couriermail, @QuestNewspapers, @QPSmedia, @abcnews, and individual journalists like @latikambourke all feature in the top 50.

More thoughts on all of this soon – and a request: what would you be interested in seeing? What do you think would best measure the role of Twitter during the disaster? Let us know in the comments.

About the Author

Dr Axel Bruns leads the QUT Social Media Research Group. He is an Associate Professor in the Creative Industries Faculty at Queensland University of Technology in Brisbane, Australia. Bruns is the author of Blogs, Wikipedia, Second Life and Beyond: From Production to Produsage (2008) and Gatewatching: Collaborative Online News Production (2005), and a co-editor of A Companion to New Media Dynamics and Uses of Blogs (2006). He is a Chief Investigator in the ARC Centre of Excellence for Creative Industries and Innovation. His research Website is at snurb.info, and he tweets as @snurb_dot_info.

Related Articles

Share

(19) Readers' Comments

  1. One thing I was interested was a simple comparison of number of tweets per hour (basically your first graph) and river height (from http://www.bom.gov.au/fwo/IDQ65389/IDQ65389.540198.plt.shtml)

  2. Hi Will,

    that was one of my first thoughts, too (and one reason I asked for historical river height data on Twitter a few hours ago – the link you’ve included only shows the last five days, so it’s moved past the 11-14 Jan. period I’m looking at here).

    Anyway, even without these data I don’t think there’s a particularly strong correlation: the standard diurnal patterns still dominate in the data. As you can see in the first graph, we have significant drops in activity during the Australian night, even on those nights where the river stayed at its peak (the highest peak was reached around 4 a.m. on 13 Jan., from memory, but Twitter activity during that time is still way below daytime levels).

    That’s not to say that there weren’t (probably) unusually high levels of activity during that time, compared to normal nights. At a later stage, we might be able to compare events like #qldfloods with long-term averages, to see how far above standard Australian nighttime levels that night was – but right now we just don’t have that baseline data for the Australian Twittersphere…

    Axel

  3. Pingback: Tweets that mention Mapping Online Publics » Blog Archive » The Queensland Floods on Twitter: A Brief First Look -- Topsy.com

  4. Yeah, I assumed that diurnal pattern would be much stronger than river height on a simple correlation – but there might be something useful if number of tweets for each hour is normalised against a broader set (getting an idea of the standard fluctuation in tweets per hour in Australia (over the 168 hours in a week) might be a useful thing for normalising data like this).

    This might show if discussion peaks before the river does (Twitter as warning), or after (Twitter as problem solving).

    On the river data, yes, I was just looking myself, but can’t find a way to get older data. The BOM may have it on their side, just not on the website.

    • Will,

      yep, the normalisation idea is what I was getting at in my comment above – but getting data on the full average volume of tweets in Australia over a random 24 hour period is tricky (not least because we’d need to know all Australian Twitter users, or at least know when we have a representative sample of them).

      Still, we’ll try to get there…

      Axel

  5. I’d love to see a breakdown of tweets to some of the major hash tags – #qldfloods and #thebigwet in particular – based on whether they were:
    - helpful information on flood levels, service availability, etc
    - fundraising info
    - well wishes
    - celebrity replies
    etc.

    A timeline on this with information on when people started thinking about fundraising and the like would also be fascinating to see.

  6. hi guys,

    great stuff. The interplay between ‘people on the ground’ vs. ‘official’ sources in distrubting accurate and timely info would be interesting to look at. Maybe you could look futher into retweets and timelines to see the flow of info from unverified to mainstream sources. Can we develop an algorithm to determine the ‘truth’ and / or build up an accurate context from a tweet?

    • mattner_d, Sandra,

      All good ideas, yes ! We’ll start exploring those possibilities – will update the blog when we’ve got anything to share…

      Axel

    • Hi Axel
      Great post thanks for the valuable information. The use of Twitter during these devastating floods was instrumental in that it brought together a community in crisis and enabled them to build a collective picture of the crisis situation as it unfolded. The convergence of social and mobile media has truly revolutionised the way we communicate and share news, although it is important to ensure the information we receive comes from a credible source. During the #qldfloods there was false and misleading information circulating around the twittersphere, although with the presence of the Queensland Police Service and other emergency management groups this false information was quickly eradicated.

      The QLD Police Service or #QPSmedia was a fantastic reliable source of information and was easily followed as each of their tweets consistently contained the #qldfloods hash tag. Twitter enabled them to not only disseminate vital communications but include links to maps, video and images through the use of their Twitter account in conjunction with their Facebook profile. Since the #qldfloods it has been amazing to see #QOSmedia’s continued use of Twitter and witness other government organisations jump on board the social media ban wagon. Thanks to you and your team for your value research in the very important field.
      Lisa

  7. Pingback: Tweets that mention Mapping Online Publics » Blog Archive » The Queensland Floods on Twitter: A Brief First Look -- Topsy.com

  8. Pingback: Social media saving lives: user generated flood relief | Reading Room Australia blog

  9. Pingback: Mapping Online Publics » Blog Archive » Media use in the #qldfloods

  10. Pingback: Mapping Online Publics » Blog Archive » Twitter Events in Perspective

  11. Pingback: The fourth essential use of social media in a crisis. « idisaster 2.0

  12. Pingback: Mapping Online Publics » Blog Archive » Image sharing in the #qldfloods

  13. Pingback: Mapping Online Publics » Blog Archive » Twitter in the Christchurch Earthquake, Pt. 1

  14. Pingback: Mapping Online Publics » Blog Archive » Social Media and Emergencies, in German(y)

  15. Pingback: Mapping Online Publics » Blog Archive » Event: Social Media in Times of Crisis

  16. Pingback: An Investigation into the Effectiveness of Social Networks in Emergency Situations - Online Conference On Networks and Communities

Leave a Reply