Data cleaning approaches in Web2.0 VGI application

Abstract

As the web provide a more flexible and sophisticated platform for information dissemination and exchanging, the idea that allow user to add and upload geospatial-related data in GIS-enabled online site to promote the geographic information sharing in public is becoming a new research topic in GIS. There are some real-world websites already partially implement VGI thought like wikimapia.org and openstreetmap.org, but considered the new characteristics of VGI data, there needs some new methods especially data cleaning approaches in the process of handling VGI data to make the non-professional user generated data into applicable information. The features of VGI data are as follows: since general users can add and change data, the stored data should be updated frequently, result in an abundant and keep updated geographic dataset; data acquired by non-professionals with non-professional equipment, so there is no guarantee of quality about the data; contributors are spontaneous, the density of data will be unpredictable, data distribution will be uneven and inconsistent. These characteristics brought challenges for VGI data management, New data cleaning methods must be used to solve problems like data clutter, . Data cleaning, including the merge, delete, correct operations on geographical features to make it into a effective and consistent form, in general, merge and delete operation can be achieved automatically, the correct operation require some human user's interaction to complete. Spatial information as well as attribute information must be taken into account in the geographic data cleaning like duplicated points finding and repeated description merging, on this basis, we will discuss several data cleaning methods for VGI data and their experiment result as well as some possible application environments in using these methods.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 14 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.