Documentation of human rights violations using heterogeneous geodata from remote sensing and (social) media

Abstract

Modern remote sensing from space, today with spatial resolutions of up to 15cm and in principle daily repetition rates, allows to document changes on the land surface with high precision. This makes it possible to record the development of armed conflicts, refugee camps, population and settlement trends in border regions, the demolition of informal slum areas and the like. Media report on these and other social issues in image and text formats. By means of natural language processing and methods of geolocalization, it is possible to evaluate the content of media reports on a spatial basis. In social media, developments on these issues are reported from a personal, subjective perspective, sometimes with a clear spatial reference. The combination of these three independent data sources holds high potential: On the one hand, reports from each source can be cross-checked and verified, and on the other hand, the limitations of individual data types or disciplines can be reduced. Remote sensing, as an example, allows to capture the physical changes of the land surface or objects on it. Nevertheless, the respective political, demographic, social or other processes behind them are not always obvious. The demolition of an informal slum area can be documented by remote sensing and helps to locate and measure these demolitions unambiguously. However, the demolition can be a planned development for rebuilding, but it can also be an eviction of illegal squatters. Media reports and/or social network discourse on this issue can be used to contextualize and capture the context of such events, to classify or better understand the forces driving these processes. We will show by some selected examples how the joint analysis of these heterogenous data sets allows a more holistic picture for a critical geography.