Automatic construction of travel itineraries using social breadcrumbs

TLDR

Vacation planning is laborious and requires many resources, yet Flickr’s geo‑temporal photo metadata can be mapped to points of interest, providing a rich source of user activity data. This study aims to automatically generate intra‑city travel itineraries by exploiting the geo‑temporal breadcrumbs left by millions of tourists on Flickr. The method extracts individual users’ photo streams to estimate locations, dwell times, and transit times, aggregates these into a POI graph, and then constructs itineraries from the graph based on POI popularity and user constraints. Crowd‑sourced evaluations on Amazon Mechanical Turk with ~450 workers show that itineraries produced from Flickr data are of high quality and comparable to professionally generated bus‑tour itineraries.

Abstract

Vacation planning is one of the frequent---but nonetheless laborious---tasks that people engage themselves with online; requiring skilled interaction with a multitude of resources. This paper constructs intra-city travel itineraries automatically by tapping a latent source reflecting geo-temporal breadcrumbs left by millions of tourists. For example, the popular rich media sharing site, Flickr, allows photos to be stamped by the time of when they were taken and be mapped to Points Of Interests (POIs) by geographical (i.e. latitude-longitude) and semantic (e.g., tags) metadata.Leveraging this information, we construct itineraries following a two-step approach. Given a city, we first extract photo streams of individual users. Each photo stream provides estimates on where the user was, how long he stayed at each place, and what was the transit time between places. In the second step, we aggregate all user photo streams into a POI graph. Itineraries are then automatically constructed from the graph based on the popularity of the POIs and subject to the user's time and destination constraints.We evaluate our approach by constructing itineraries for several major cities and comparing them, through a crowd-sourcing marketplace (Amazon Mechanical Turk), against itineraries constructed from popular bus tours that are professionally generated. Our extensive survey-based user studies over about 450 workers on AMT indicate that high quality itineraries can be automatically constructed from Flickr data.