Life and death of volunteered geographic information contributors in a large online community - the case of OpenStreetMap

The advent of the Web 2.0 has democratized both the production and dissemination of knowledge by enabling communities of online contributors to generate content collectively. This thesis focuses on “Volunteered Geographic Information” (VGI), a type of user-generated content (UGC) oriented toward geographic information. The provided content is known to be highly heterogeneous in coverage, nature and quality, reflecting a patchwork of motivations, interests, knowledge and skills of individual contributors. Characterizing VGI data requires understanding contributors’ behaviour. Typologies of contributors are proposed in an attempt to link VGI contributors with the nature of the data they provide. Those typologies are directly or indirectly related to the time spent by the contributors in a project, but they do not use a formal temporal perspective to understand their behaviour. We considered the time spent by contributors in a given VGI project as an essential component for understanding their contribution patterns (e.g. volume, content, quality). In order to fill this knowledge gap regarding how the time in the project may have impacted contributors’ behaviors, I analyzed the behaviour of the OpenStreetMap (OSM) contributors, of a large VGI community. I identified different events that affected enrollments and withdrawals over a project’s history using time series analyses. I established the phases of contributors’ life cycle using survival analyses and linked their contributions to the different phases. Six distinct phases were identified in the life cycle of OSM contributors. Analyses revealed that these phases were grouped into three major stages: An “Assessment” stage that last a few months, followed by an “Engagement” stage that can extend over more than a decade, to eventually move to a “Detachment” stage over which the contributors leave the project. Analysis of contributions at each phase revealed that contributors’ behaviour is dominated by two distinct processes. When contributors enroll in a project, they seem to be driven by a learning-adaptation-dominated process before switching to a cumulative-damage-dominated process followed by a withdrawal from the project. In parallel, I found that the diffusion of innovation theory (DoIT) had an important impact all along the project’s history. This research not only shed light on online contributions but also reveals different aspects of human behaviours