So today’s lecture focused on social media and elections.
Here’s a summary of what my Facebook friends were talking about on election night:
50% the AFL
40% ‘we’re all going to hell in a hand basket’ type statuses (Labor/Greens voters)
1% comments about why it is a good thing Liberals got in, or, ‘I don’t care’ type statuses
9% other things, because it’s a Saturday night
This information is fairly useless in terms of getting an idea of what the ‘general public’ (god I hate that phrase) wants because my Facebook friends list is not an accurate sample group, because, well, if I’m friends with someone its usually because they have pretty similar views to myself.
So continuing on with this, how can we know that comments on social media are an accurate idea of what the ‘general public’ wants? when:
– Not everyone has social media or even the internet
– Only about 1% of people are active sharers.
There’s some information about this here: http://www.smh.com.au/digital-life/digital-life-news/facebook-checked-by-9-million-australians-every-day-20130820-2s7wo.html
As you can see, 40% of Australians check Facebook every day, which is a lot, but it certainly isn’t everyone. On a similar topic, here are some stats from a few months ago which I found quite interesting:
From an assessment of my Facebook friends we can see that they would much rather discuss the AFL than politics, AFL may not impact on our lives as greatly as which political party will win the election, but it’s much more exciting to watch. Go Geelong!
The lecture also covered the topic of wordclouds, which is something I’m much more interested in than elections, and besides being pretty and fun, I think they are an interesting way of arranging data.
Here’s a wordcloud I made for this very blog!
There are two things that this wordcloud made me think:
1. I’m concerned that the word ‘something’ is used so often.
2. Who is ‘don’?
I think wordclouds are useful, as you can tell it makes sense that the words ‘social’, ‘media’, ‘future’, ‘people’ are all used in this blog, but obviously it doesn’t give you an idea of how they are being used. So it is useful but I guess it can’t be used in isolation as a tool of gauging opinion. I think in terms of real-world applications, wordclouds are something that I would find quite useful. It gives you an idea of what people think and it also is presented in a professional looking manner, so now I thought it might be useful to run through the steps involved in creating one.
This is the site I used: http://www.wordle.net/. There are others but I found this one to be the most popular and I also found it easy to use.
So for this wordcloud, I am going to analyse the words I use most in my tweets.
Step 1. Go to Twitter, go to settings, click on ‘request your archive’
Step 2. Check your e-mail where the archive will be sent to. It says ‘please be patient’, but it only took a minute for me to receive it.
Step 3. Click ‘go now’ and ‘download’
Step 4. (This step is just for fun). Before I make the actual wordcloud I thought it would be fun to guess my most used words. I guess ‘I’, ‘Tweet’ and maybe ‘Penguin’ as a random one.
Step 5. Go into the excel sheet you’ve just downloaded containing your tweets, copy them into http://www.wordle.net/create, (if you Tweet a lot of links, you will have to go through and delete them but it should only take 5 minutes) click ‘go.’
Step 6. Choose how you want the words displayed, the colours, basically just make it look good.
Here it is!
I guess ‘I’ doesn’t count as a word. Or ‘Tweet’. and ‘Turtle’ and ‘Dugong’ were more popular than ‘Penguin.’ Ah well. But that does go to show that wordclouds can show insights that you didn’t expect.
You can see from this that there are real-life applications for using wordclouds. For example, I know in advertising it could be an interesting way to display consumer insights, and if you were going for a job, it could be a good way to show what you talk about and are interested in.
So to conclude, wordclouds are much more exciting than elections.
Yep, that’s what I learnt this week