Virtual City Explorer: A crowdsourching tool to locate and describe static points of interest in cities.

LoadingLoading previews...
wais-seminar-20181212.mp4
Video
Download (919MB)
    wais-seminar-20181212.mp4
    wais-seminar-20181212.mp4
    B233
    B233
    2 files in this resource

    Virtual City Explorer: A crowdsourching tool to locate and describe static points of interest in cities.

    Abstract: A common issue among European municipalities is the lack of information of mobility infrastructure Points of Interest (PoIs). This information is valuable both for powering services to citizens and as a mean to audit the urban health and environment of the city. Currently, municipalities need to send their employees to the field to do the counting, an expensive and error-prone approach that does not scale in the size of the area to be covered. Alternatively, one can rely on Volunteered Geographical Information (VGI) systems to crowdsource the data, but at the expense of having no control over data updates. We propose a faster and cheaper solution to tackle the problem by taking advantage of virtual imagery to use paid crowdworkers who can perform the item locating task remotely with no need of to be physically in place and without having any prior local knowledge of the area for which the exploration is required. We implemented a standalone crowdsourcing system named Virtual City Explorer (VCE) which allows collecting locations and images of PoIs on virtual spaces with paid crowdworkers. Our system takes as input a virtual space (e.g. a Google Street View instance), a type of PoI and an area of interest (defined as a geo-spatial polygon) and returns coordinates of instances of the target PoI type inside the area of interest discovered by a fixed, configurable, number of crowdworkers. Each crowdworker is asked to explore the area of interest finding a (configurable) number of PoIs, being rewarded upon completion. Our first experiments resulted in being very encouraging, showing how the VCE can effectively be used to find and collect locations of PoIs, in a cheaper and faster way. Speaker information Eddy Maddalena is a research fellow in the Web and Internet Science (WAIS) group of the University of Southampton, United Kingdom. Eddy got a PhD in Computer Science in 2017 at the University of Udine, Italy. During his PhD, Eddy mainly focused on Information Retrieval (IR) and crowdsourcing to create human annotated test collection of documents to be used to evaluate the effectiveness of IR systems. After holding his PhD, Eddy moved to the University of Southampton where he is the leader of the Crowdsourcing Work Package for the H2020 QROWD Project, and designs and develops crowdsourcing based solutions to improve mobility and reduce transportation issues for smart cities, which aim to include “human in the loop” participation in the Big Data Value Chain information flow.

    Actions (login required)

    View Item View Item

    Toolbox

    There are no actions available for this resource.