Generating Personalized Spatial Analogies for Distances and Areas

1) scales2015_master_final.csv
This is a file of containing information of all landmarks that we use in the project.
id: id that we assigned
name: name of landmark
latitude: latitude of the landmark
longitude: longitude of the landmark
city: city that the landmark is located
state: state that the landmark is located
flickrCount: total count of photos from flickr
areaLandmark: 1 - area landmark 0 - non area landmark
osmid: assigned OpenStreetMap id if the landmark is areaLandmark. If you want to get polygon information, you can use this id to get the polygon by using OSM api(

2) results_energy_functions.pdf
In this file, we presented top five personalized spatial analogies produced by our approach for seven university locations (University of Washington, University of California, Berkeley, Columbia University, Massachusetts Institute of Technology, Carnegie Mellon University, University of Michigan, Georgia Institute of Technology) listing a set of distance and area targets respectively. We generated tables for a range of distance and area inputs and users' locations, but where each term in the energy function is turned off at a time. 

Yea-Seul Kim / iSchool / University of Washington /
Jessica Hullman / iSchool / University of Washington /
Maneesh Agrawala / Computer Science Department / Stanford University /

Please send an email to if you have any question.