This year Twitter made its impact felt on the Party Conference circuit. I went to NESTA Fringe meetings in Brighton last week and in Manchester this week. I came up with the idea of seeing whether the Tweetstreams might tell us something about the relative states of those two parties. The results are, I think, intriguing.
#cpc09 beats #lab09 hands down
I’ve looked at the main hashtags being used at both the Labour and Conservative Party conferences over a comparable period in each case (beginning on Sunday and ending on Thursday). The total volume of Tweets with the #lab09 tag was 10,379 compared with 12,733 with the cpc09 tag. The numbers are derived from time series data kindly provided by What The Hashtag?! and I acknowledge the help of Mark Bockenstedt for his advice in understanding how to use the API.
Naturally, all I’m doing is looking at Tweets tagged with those particular hastags – I don’t know at this stage what the contents or stance of the Tweets might be; whether positive or negative.
Time series data . . . hmmmm
We can also examine the flow of Tweets over time (and looking at the structure of flow is always instructive).
This picture shows the daily volume of Tweets with the #lab09 tag, beginning on the Sunday (Day 1) and running until Thursday (Day 5). When I saw this picture, I wasn’t particularly surprised – it shows a build-up of activity each day with a ‘peak’ on Tuesday; the day when Gordon Brown did his ‘big speech’ to conference. Looks like activity diminished somewhat on Wednesday and Thursday – and indeed, recalling the news coverage at the time there was talk of the conference ‘going a bit flat’.
Looking at the daily volume of Tweets tagged #cpc09 is a bit more surprising. Day 2 (Monday) was the day Boris Johnson ‘did his thing’ and William Hague gave a keynote. Day 3 was George Osbourne’s Gloomy Day. Day 5 (yesterday) was David Cameron’s ‘big speech’.
Now examine the two charts together
Look at the volumes; only once did the #lab09 tag reach over 3,000 per day; and that was when Gordon Brown spoke. And the daily volumes were consistently larger for #cpc09. Activity levels higher across the piece. And it seems to me, by observation, there seems to be more ‘momentum’ in the #cpc09 hashtag. Certainly, I noted (and Audioboo’d about) the generally less cheerful and relatively more cheerful feels of the Labour conference people I observed versus the Conservative conference people. Note I was just in each city (actually on the Tuesday) and at Fringe events outside the ‘security zone’.
It’s just an observation – and you may have some thoughts
What does it mean? I should add that I’ve looked at hashtag activity extending both sides of the Labour conference and for the run up to the Conservative one. There are no glaringly obvious patterns and, in any event I have produced comparable stats on the same chart scales for each conference.
I’m still thinking about what, if anything, this analysis tells us. It could be that the volume of #cpc09 tweets reflects relatively more negative traffic (ie Labour supporters using the hashtag to criticise the Tory conference) than is in the #lab09 Tweetstream. It could just be that the disquiet of the Labour supporters is reflected in their lower use of the #lab09 hashtag – staying quiet rather than Tweeting negative thoughts. Please feel free to comment if you have further thoughts.
Why no #ldc09?
Well, I can’t get back in time as far as the Liberal Democrat conference – Twittersearch says “No older Tweets available” and they’re not there on WTHashtag?! either. Which brings me to a further thought – Tweets are ephemera. they vanish into the ether after about 10 days or so as I understand it. However their nature Which, will, I think, become an issue if Twitter starts to have an impact on the political process . . .
Because it should have been #ldconf
Thanks to Tory Bear for pointing out my error, and also, see his comment below. My reading (such as it is) of the Tweetstream from #cpc09 does tally with his view.
So here we have the Tweets from the Liberal Democrat conference alongside the Labour and Tory ones (it’s not quite midnight oil burning yet . . ).
The choice of hashtag is a bit odd, I think – not including the year does not follow ‘best emerging practice’ such as it is. I wonder, to what extent, the choice of this hashtag was really planned?
Now for the timeseries data, shown on the same scale as #lab09 and cpc09 above. At this scale, the detail is not apparent – and when I look at the numbers, the daily volumes vary from about 700 Tweets per day (Shall I define a new unit – Tpd?) to 900 Tpd. So the idea tha the Lib Dems are somehow more sociable and chatty does not seem to be borne out by the evidence.
Endnote: You can find my other material on the conference fringe events on the Amplified09 website:
Amplifying Nesta at the Labour Party Conference
Amplifying NESTA at the Conservative Party Conference – Getting Creative
the LibvDem hashtag was #ldconf
Labour are trying that the only reason the tories got more tweets is that they were being negative. A read through them does not reflect that. yes there were some negative comments but the majority were pro comments from the blue team.
also depends if a hashtag trends, that tends to encourage an awful lot of spam…
… also twitter insists it is saving all tweets. somewhere.
maybe they will sell them to historians? market forces. It is possible to save an entire hashtag as Will did for #tal09. @willperrin tweeted the link: @cyberdoyle @rayduff go http://bit.ly/qgaYK and follow the instructions to create free doc of all tweets on a tag in a date range
Maybe each conf should do this for its own hashtag, and save its own history?
Thanks for the comment Chris – yes, I haven’t filtered the Tweetstreams at all – so there will be hashtag spam in there.
I’m also thinking about how to do content analysis; it would need to be automated in some way.
Brian
Thanks Tory Bear.
I have sampled both #lab09 and #cpc09 tweets (though not on a scientific basis) and I can’t find any particular differences in the levels of positivity or negativity between the streams. Obviously both streams have critical and sometimes ridiculous negatives. And you can see some obvious Labour supporters making critical comments on #cpc09; likewise some obvious Tory supporters making negative comments on #lab09.
I argue, that to a first approximation, the levels of negative comment in each direction between the parties are likely to be about the same.
As far as negative comments from supporters – ie the Blue team being negative on #cpc09 and the Red team being negative on #lab09 – I don’t see where the evidence is for that. Rather, I think the timeseries data supports the observation that the Labour conference ‘went a bit flat’ in its last 2 days.
My own (admittedly unscientific) observations are in the Boos linked to above.
Brian