I've been waiting to see the content we have so far (1000+ consumer wine reviews) in a workable data format. It'll take some time before we see the full "big data" power of Real Time Wine interactions - but for the moment, we can already see some interesting trends arising from light semantic analysis of the review text.
Check it. A word cloud (the larger the word, the more frequent its use) of the first 250 reviews:
To be fair, the first 200 or so reviews were mostly by me - but I love visualisations like this. Words popping out:
- like - very consumer orientated, I like this, I don't like this
- fruit, fruity
- smells, tastes - have always said a good structure for a review is "smells like, tastes like, feels like"
- light, spicy
- berries, chocolate, creamy - I want to explore this in more depth, I wonder if people might search for a taste they like and use that as a route into finding recommendations
- nice - Mrs Jones down the road is niiiice.
All fairly easy to understand consumer-speak. The trick is, can we make this more valuable than the Platters reviews which serve the top end of the market (and require some intense knowledge to understand)?
What's needed from a linguistic sense to supplement and encourage peer recommendation?
We shall see! More data shortly...