Assessing Crowdsourced POI Quality: Combining Methods Based on Reference Data, History, and Spatial Relations

Abstract

Abstract:With the development of location-aware devices and the success and high use of Web 2.0 techniques, citizens are able to act as sensors by contributing geographic information. In this context, data quality is an important aspect that should be taken into account when using this source of data for different purposes. The goal of the paper is to analyze the quality of crowdsourced data and to study its evolution over time. We propose two types of approaches: (1) use the intrinsic characteristics of the crowdsourced datasets; or (2) evaluate crowdsourced Points of Interest (POIs) using external datasets (i.e., authoritative reference or other crowdsourced datasets), and two different methods for each approach. The potential of the combination of these approaches is then demonstrated, to overcome the limitations associated with each individual method. In this paper, we focus on POIs and places coming from the very successful crowdsourcing project: OpenStreetMap. The results show that the proposed approaches are complementary in assessing data quality. The positive results obtained for data matching show that the analysis of data quality through automatic data matching is possible but considerable effort and attention are needed for schema matching given the heterogeneity of OSM and the representation of authoritative datasets. For the features studied, it can be noted that change over time is sometimes due to disagreements between contributors, but in most cases the change improves the quality of the data.

Graph-Based Matching of Points-of-Interest from Collaborative Geo-Datasets

“…require larger distance thresholds than others (e.g., cafés, bars, shops), the general practice seems to be to establish a universal threshold based on the Euclidean distance of pairs of matching POIs from a training set. Threshold values ranging from 60 up to 1000 m can be found in the literature [2,[26][27][28].…”

“…As mentioned above, previous works report that corresponding POIs are sometimes found hundreds of meters from each other. This greatly limits the effectiveness of considering the topological similarity POIs, as suggested by [2,29,30]. This type of similarity considers, for example, whether two POIs from different sources are located inside the same building footprint or at the same side of the street.…”

“…The simple unweighted sum [30,31] as well as decision-trees [37], logistic regression [26], entropy-based [27] and belief theory [2,38] methods have been tried in the past.…”

“…Although authoritative and proprietary geo-datasets usually detain high levels of data correctness and consistency, they frequently lack semantic information unrelated to the specific administrative and commercial purposes they serve. Volunteered Geographic Information (VGI) platforms, on the other hand, provide freely accessible information about urban features to a considerable degree of geometric accuracy and semantic comprehensiveness [1,2]. However, this information is not available in one specific VGI platform, but is instead dispersed across different platforms, each of which with its purposes, strengths, limitations and community of volunteers.…”

“…As is frequently the case, these two research areas intersect when data matching is a step of the quality assessment workflow [1][2][3][4][5]. For example, finding the same feature at different datasets increases the degree of trust that this feature exists in the real world, whereas comparing and merging their attributes serves respectively to evaluate and improve their semantic accuracy and completeness.…”