SpaceRef
SpaceRef is a corpus of street-level geographic descriptions. Pedestrians are walking a route in an urban environment, describing their actions in realtime. Their position is automatically logged, their speech is manually transcribed, and their references to objects are manually annotated with respect to a crowdsourced geographic database. The corpus is freely available for research purposes. If you download it, please let us know, describing shortly what you plan to use it for or why you downloaded it. Please refer to the publication [GötzeBoyeLREC] below to cite SpaceRef.
Data
Files
The data consists of the following files:
- Annotated data files for each study participant
- Logs of GPS positions
- The OpenStreetMap snapshot of the study are, taken at the time of the data collection
- A file containing mappings of Openstreetmap IDs to streets
- The task description as given to the participants
- Answers from the questionnaires given to the pedestrians after their participation (to come)
- Python scripts that process OSM and data xml (to come)
Documentation
Download
Publications
- [GötzeBoyeLREC] Corpus description:
SpaceRef: A Corpus of Street-Level Geographic Descriptions
(pdf, bib, poster)
J. Götze, J. Boye
Proceedings of the Tenth International Conference on Language Resources and Evaluation (LREC) 2016.
- [GötzeBoyeLBS] Salience models for landmarks in route directions:
Learning Landmark Salience Models from Users' Route Instructions
(pdf)
J. Götze, J. Boye
Journal of Location Based Services, 2016.
Contact