Inferring attitudinal spaces in social networks

Ideological scaling methods have shown that behavioral traces in social platforms can be used to mine opinions at a massive scale. Current exploit one-dimensional left–right opinion scales, best suited for two-party socio-political systems and binary divides such as those observed the US. In this article, we introduce new method overcome limitations of existing by producing multidimensional network embeddings align them with referential attitudinal few nodes. This allows us infer larger set dimensions from graphs, embedding users spaces where stand indicators several including (in addition cleavages) attitudes towards elites, or ecology among many other issues. Our does not rely on text data is thus language-independent. We illustrate approach Twitter follower network. Finally, show how our analyze shared within various communities networks. analyses extreme political are also more homogeneous ideologically.