Aims to quickly create a pub crawl from 2 given locations. The back-end is written in Java. The front-end is in JSP, Javascript and some javascript google maps stuff. All dev was done in eclipse.
-
Given 2 geo-locations (a, b) it will first create an area where a and b become the 2 opposite corners to form a rectangle.
-
This rectangular space is the area we look for pubs and clubs (places). Using Yelp & Qype as data sources it downloads all the places in the area.
-
The duplicate places are aggregated together so as to not offer duplicate places.
-
Crappy places and ones that aren't reviewed are removed, if possible.
-
Using a mixture of distance travelled, reviews & number of reviews [bayes] it will generate an optimal crawl.
-
Google maps displays the pub crawl.
-
You cannot create a pub crawl that goes in a circle. It only works if there is a start and end location and aims to create a straight line or near curved line.
-
The weightings given for (reviews and distance) work Ok for a city but fall down when trying to create a pub crawl in the sticks.
-
The distance between points (geo-located) does not take into account the curvature of the earth. e.g. 1 latitude degree becomes less and less (in terms of actual, physical distance) the further north you are.
-
The 'website' is pap. It's my first attempt at a website and it needs completely rewriting.
-
Bad data. Can't really do anything about idiot users who think Nando's comes under nightlife/drinking. It always gets highly rated and therefore appears in pub crawls, sigh.