We’ve already seen Dish.fm crunch the massive amount of data from online restaurant reviews to recommend the best eateries for its users, but our latest spotting aims to go even further to personalize results. Nara is an app that uses a Pandora-like neural network that learns users’ tastes every time they use it and offer accurate recommendations for eating out.
Created by a team consisting of “neuroscientists, computer scientists, astrophysicists, artists and entrepreneurs”, the app first asks users a few questions about the kind of eateries they like, based on food types, atmosphere and demographic, among others. When a suggestion is made, it can either be upvoted or downvoted depending on the user’s opinion, which can be logged either before or after they’ve tried it out. Foursquare check-ins are integrated to keep track of where users have been before, and which restaurants they like to go to regularly. By checking this against the decisions made by every other Nara user, the system quickly begins to intuit the kinds of decisions made by those with similar tastes. The following video explains more about the app:
The app is available for free on the App Store and Google Play. While Nara is currently concentrating on the restaurant vertical, reports suggest it is looking to move into doing the same for other fields, such as the hotel sector. Rather than viewing this as a separate endeavour, the data will feed into the same neural network that provides restaurant matches. The end goal will be a service that could recommend a bar based on users’ previous engagements with totally different businesses, or even the kind of music they like – painting a much broader picture of each consumer. Could this be the future of finding new brands on the web?
No comments:
Post a Comment