If you have been on Yelp and other sites that review places or things, like Amazon product reviews, don’t you hate the users who determine their rating points based on a minuscule aspect of the restaurant, product, or whatever? Like for example, people who give restaurants with amazing food one star simply because they couldn’t find parking? Or they slammed a kitchen appliance because it didn’t come in a color that they liked. One wonders if there was some way to develop an algorithm that could remove the scoring on reviews that slammed something due to a non-core attribute being rated low. I guess that is one of the issues with crowdsourced ratings and reviews; how do you keep the quality level of a rating consistent. If I write a review, it’s usually based on promoting a place/product that I’m going to/using to others or dissuading them from going to/purchasing it. But as I look over both Yelp and Amazon reviews, especially the negative ones, they are rife with ratings crushing complaints about minor aspects – at least in my opinion.
I’m guessing that this is not a trivial problem – unless you can map the data points and give the user a specific set of attributes to rate on, how can you get a clear sense of the ratings of a place/thing with all of the extraneous mumbo-jumbo removed. Something to think about – maybe a startup focused on scrubbing the set’s unimportant data would be interesting.