Introduction

Sloyka documentation Date: June, 2024 Version: 0.6 SOIKA is a library aimed at enriching digital city models with data obtained from textual data of citizens“ digital footprints, as well as at modeling vernacular assessment of urban environment quality.

Its main element is a constructible spatial semantic hypergraph, augmented by machine recognition of urban entities and locations.

The SOIKA’s final result is a spatial semantic hypergraph, which generates after two main stages: data receiving (messages from the social network, mentioning particular city objects in them) and additional processes of data tagging of the collected data to obtain new columns in the resulting GeoDataFrame. The resulting hypergraph can be used to predict events within existing urban objects (module Regional activity), or to visualize already existing nodes and links and their further interpretation (module Graph visualization)

SOIKA also provides methods for modeling social risks regarding the emotional evaluation of mentioned places.

Main features

  • Social media parsing: getting posts, comments and replys

  • City services and places extraction

  • Emotion and text classifiers categorizing

  • City’s topic modelling

  • Spatial-semantic graph building

  • Regional activity evaluation

SOIKA’s Community chat: https://t.me/soika_community