By Liqiang Cheng, Naiqi Jin and Jason Yap
Project Description Project summary: A Geo-spatial search system that collects and combines data from various sources into a single searchable repository. Goals: Collect data about locations of interest from multiple sources. Provide a User Interface that allows a user to query a geographic area for interesting locations. Satisfy queries efficiently (results returned in real time).
Translated Index Query Architecture Crawler Urbanspoon.com Yelp.com External Sources Data Repository and Spatial Indexes User (Web Browser) User QueryResults Result Set Crawled Results UploadQuery Update DBMS & Index Web Server
Crawler yelp.com urbanspoon.com
Crawler two data sources yelp.com urbanspoon.com crawling several postal codes around USC information crawled name address URL latitude, longitude average review (1-100 points) number of reviews
Example Crawled Results yelp.com
Indexing and Retrieval MySQL Spatial Index was used to index locations for efficient querying. MyISAM tables specified with “SPATIAL INDEX” create an R-tree index which supports efficient range queries. GPS coordinates for each location was extracted by the crawler and registered with the DBMS and index through the Upload component. To generate a query for retrieval, a user specifies points of a polygon which translate to GPS coordinates. A Minimum Bounding Rectangle(MBR) specifies the area for the DBMS query in SQL.
User Interface Advertisement module - Commercial scrolling advertisement and propagation such as facebook and twitter. Geo-spatial Search module (Primary UI) - Conducting search query by drawing polygon. - Google maps presentation ordered by Name, Rating, Type or Votes No. (Different type has different color indexed Marker) Advertise ment Geo- spatial Search UI (Web Browser)
User Interface
OrderBy Number of Votes OrderBy Average Rating Facebook popularization OrderBy Average Rating
Demo Website: