(This article was originally published at Data Mining: Text Mining, Visualization and Social Media, and syndicated at StatsBlogs.)
When I pushed out the first version of Beach G33k / Beach Ball (I'm not actually sure what the thing is called) - a search engine for beaches - I did as close to nothing in terms of features as possible. This approach to release follows the release early, release often paradigm which leverages user feedback and small iterations to help guide a service according to the user needs.
Thanks to a comment by Viliam Kanis I've now pushed out an update which does use some amount of stemming to improve searches. Thus, if you search for 'snorkeling' you will also find beaches with associated content which mentions things like 'snorkel' and 'snorkeled'.
Below is an example of a search result for 'snorkeling' which surfaces Lydgate Beach Park in Kauai.
If you are heading for Kauai or Maui in the near future, take the site for a spin.
Please comment on the article here: Data Mining: Text Mining, Visualization and Social Media