Cultural Trends in Hansard: Looking at Changes, 1994-2012

Continuing on my previous work, where I used Google Trends to see whether we could detect the “Warrior Nation” thesis in Canadian search habits, I wondered if we could approach this simple problem in another direction. Instead of the grassroots perspective, what was happening in parliament? What would we find? While I’m not a political scientist, nor a political historian (although I’m the Political History Group’s webmaster!), I thought this might be a useful problem to quickly tackle tonight.

The federal government has its transcripts of debates available since 1994. Using Mathematica, I decided to scrape the full text of this material. The resulting amount information was Big: 800MB of plain text files, an amount that even my 16GB system struggles with working with. So how can we make sense of this sort of big data, in a meaningful way? Two tricks: first, topic modelling; secondly, n-grams (dealt with primarily in a later post).

Topic Modelling Hansard

Topic modelling works quite well with this sort of material (although this size of a corpus is why we put the ‘working with large files’ section into the Programming Historian 2 article). There has been quite a bit of thought put into the limitations of topic modelling, but as a start it does show historical change over time within a very large corpus. After putting the material into MALLET, and subsequently into our topic model visualizations, the results are interesting. Click here if you want to see the full list of material.

One topic immediately jumps out:
international canada peace mr nato war world peacekeeping conflict troops nations united people kosovo situation humanitarian foreign role genocide
This is unsurprising, in the Warrior Nation thesis. There is a noticeable drop off in this peacekeeping, peace, etc. topic after the Conservative election in early 2006; however, the 9/11 attacks could also be seen as a significant fulcrum. We also do continue to see spikes.

Future work will have to look into where these spikes are coming from: are they coming from Liberal members? Or government members? Luckily, we can eventually automate this so we pull up the top 20, 30, or even 40 debate areas per topic.

Another topic might be relevant as well:
afghanistan mission canada canadian afghan mr minister government troops military security women defence forces international soldiers development motion support

Here, we see a topic directly related to Afghanistan, albeit defence more generally as well (we see some brief squiggles in the 1990s, but it accelerates in first early 2001 with a news spike about the Taliban and then with the height of Canadian involvement in the Afghanistan war. This topic, however, continues. If we map the two topics together, we can see the transition as well:

There is a transition, but whether we can ascribe this to the Warrior Nation hypothesis is realistically beyond the scope of this very preliminary inquiry!

A few other topics are worth looking at in general:

The general scaffolding of parliamentary business, a constant:
committee mr report standing important parliamentary speaker work secretary process house issue recommendations review national made ensure information forward

The shifting rhetoric on the budget from the blue topic
budget government billion year finance canadians debt minister spending tax years deficit cuts money fiscal care canada health cut
which noticeably declines after 2006, replaced with a new one relating to Canada’s economic action plan,
economic budget jobs economy canada tax plan mr canadian canadians government measures action businesses support credit world finance crisis
. They’re of equal importance, but do fairly neatly spell each other off.

Finally, a few other one-offs. We always seem to care about the children (won’t somebody think of them!!):
criminal code police sexual children offence mr law child person offences pornography justice dna age defence sex protect arrest

And heritage, sadly, seems to be on the decline. This may not bode well for my own budgets…
canadian cultural heritage canada culture flag canadians minister industry country mr arts national department world museums film artists quebec

… unless I switch over into remembrance!
veterans war affairs canadian service mr benefits day world men services support speaker member country forces remembrance committee served

One thing I’ll leave you with might be worth exploring in a later post, that of n-gram frequency. Here’s something to leave you with – our federal parliament does seem to talk about peace a whole lot less frequently:

Leave a Reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s