Skip to content

Home

Jump to bottom
recalegari edited this page Apr 6, 2018 · 19 revisions

In this tutorial we explain how to modify the existing GWAP-enabler to build a new GWAP that aims to classify OpenStreetMap data on the basis of a given dimension or attribute. The idea behind this tutorial is to show, on the one hand, a guided example of a GWAP-enabler implementation and, on the other hand, to demonstrate how this enabler could be used and adapted to build application for Linked Data Refinement. In particular we want to build a crowdsourcing web application to collect data for enriching and refining open data collections such as OpenStreetMap. By adopting this GWAP-enabler solution, you can build a crowdsourcing tool in which users are motivated to participate through fun incentives. The results of the classification tasks performed in the game are published as RDF triples in a JSON-LD file, guaranteeing the reusability of the results in a simple and standard format.

The use case of this tutorial is about the classification of OpenStreetMap restaurants on the basis of their type of cuisine. After having collected all the restaurants with a valid name in the Milano area, we take all the restaurants with the type of cuisine specified as Ground Truth, to build the reference gold standard to assess users contributions. All the remaining restaurants are the target of our classification process.

The main flow of the application we are going to build through this tutorial is outlined by the screenshots below. The application can be set up to be played both by anonymous users and by registered ones (using social media accounts). Users are asked to select the type of cuisine of a restaurant among a set of predefined categories. The information shown to the user are the name of the restaurants and their position marked on a map that can be browsed and zoomed. At the end of each round the score gained in the round and the position in the leaderboard are displayed to make the game more challenging.

In the following pages we will explain:

  1. the steps to download and set up a new game
  2. the process of data collection and processing to generate the final dataset
  3. some changes to the enabler in order to add further information and functionalities to the game.
Clone this wiki locally