I'm going to throw another league into the mix, sorry
My company built a social analytics, deep listening, influencer identification and influencer campaign platform for a client earlier in the year. It's solid and has a slightly unique thing about it.
It listens in across social media outlets (FB, Twitter, Google/YouTube, Instagram and Vimeo (irrelevant to the contestants)).
It checks the top level stuff such as profile "audience" (FB Page likes, followers), "interaction" (retweets, comments, tweets, content likes) and also runs a deep listening mechanism that looks for public discussion such as mentions, hashtags etc.
It also looks at audience patterns and gives weight via a gamification engine to engagement patterns that indicate an invested audience and provides a "popularity" score. That's not dissimular to the idea behind the Klout score (if anyones heard of that)... but for "popularity".
I decided to plug the contestants in just before JH to see if there is (or is not) a correlation between social media popularity and the result.
It managed to predict Hannah/Miss Dynamix and Tamera/KR in the bottom two. Wk1 and 2 I guess it got but that's hardly worth considering as there was a lower block that were all drifting around together.
Last night (Saturday) it had (lowest to highest):
Hannah
Abi
Tamera
luke
rough copy
Sam B
Sam C
Nic (way out ahead)
Today (Sunday, 11:21am) it has (lowest to highest):
Score is a normalised popularity score, similar in concept to the Klout score. The order is the same but the scores have tightened into three blocks plus Nicholas who is basically in his own little league and is almost as popular as erm... some smaller national brands, but not Mashable/Forbes/TechCrunch popularity. None of them are really brand popular.
Hannah: 751
Abi: 912
Tamera: 15,327
Luke Friend: 29,959
Rough Copy: 32,479
Sam B: 101,838
Sam C: 103,560
Nicholas: 213,456
Note (Sam C): His audience to engagement ratio is poor so his score could be grossly misleading. In other words, his metrics show that he has a smaller core audience (who are significant, but nowhere near as big as his overall score would indicate) who interact rampantly, the bulk of interaction being sharing rather than discussion. others such as Sam B, Luke and Abi provoke discussion which is more valuable.
Another is RC, they have the same pattern going on.
Tamera may be subject to massive multi-voting due to her B2 last weekend so her score may not represent her and could if it correlates again pull RC or Luke into the bottom 3 area.
Luke and Abi have a slightly unique engagement trend, most of them boost through Sat and Sun (till the afternoon) and then calm, they tend to get a kick in across the afternoon.
If Abi and Luke do that again, they could fly up the league. Abi for example was near the bottom on Sunday morning but was 4th from top come the result show. Tamera hovered at 15k and didn't pick up till 6/7pm and reached 32k
Last weekend they reached (approx):
Nic 280k
Sam C 220k
Sam B 250k
Abi 90k
Luke: 80k
RC: 70k
Hannah: 40k
KR: 35k
Tamera: 35k
Make of that what you will, I'll make my prediction later, hopefully I can get a hat-trick but I'm not as confident as there are a few unknown quantities this week (i.e. Tameras rebound vote, RC and Sam C's ratios).
Also you can't guarantee that social popularity will translate into votes.