Book recommendation is a tricky business. Asking someone directly what kind of literature they like can lead you down a pointlessly narrow path, with very little in the way of inspiration or suggestions. Over at Northwestern University, the Knight Lab's latest project aims to offer an alternative method of discovering new books.

Simply enter your Twitter handle into BookRx and, based on frequently used words, hashtags and @s, it compiles a comprehensive list of titles to suit you, as well as an equally interesting list of the words that led it to its conclusion. Offering a service that gleans information from the 'stream of consciousness' aspect of Twitter seems like a more accurate method than simply relying on direct user input.

Conveniently it doesn't require you to divulge any access information, which, when compared to the deluge of personally invasive apps on Facebook, is quite refreshing.