In the context of crisis discourses, corpus linguistic analysis of language use patterns are worthwhile in order to understand how crises are linguistically bound (Bubenhofer, 2009; Kremer and Walker, 2023). This concerns acute crises such as diseases (Semino et al., 2004) or pandemics (Kremer and Felgenhauer, 2022), but also everyday experiences of unsaftey in urban spaces (Moura de Souza et al., 2022). Metaphors of fighting the crisis (Semino, 2021) imply a search for places of the crisis (Brinks and Ibert, 2020) in order to identify, control and combat effects and causes (Chapman and Miller, 2020). When it comes to the question of experienced unsafety in everyday life, it is known from conceptual social sciences that they are modelled best as fluid spaces (Redepenning et al., 2010), which reflect forms of mobility as well as constantly changing situations (Moura de Souza et al., 2022). Even when the virus had already penetrated the population during the COVID-19 pandemic, the question of a spatial origin, the "hotspots", still played a decisive role in the search for causes (Kremer and Felgenhauer, 2022). Interestingly, however, media repertoires have developed in everyday discourses on unsafety that address this need non-cartographically in a mixture of social media and classic unidirectional TV broadcasts (Moura de Souza et al., 2022) and thus may represent a more suitable data basis for the analysis of unsafety than their mere visualization bound to map-based data.

In the search for shared "imaginaries" (Taylor, 2004), the dominant guiding metaphors (Lakoff and Johnson, 2003) and narratives (Viehöver, 2001) of everyday places, focus in Digital Spatial Humanities is directed towards multimodal analysis of imagery (Rose, 2001) and text data (Mayring, 2016). Supported by digital analysis methods, certain structural findings are available for the first time (Jannidis et al., 2017; Moretti, 2013). In the tradition of critical data studies (Dalton and Thatcher, 2014; Kitchin and Lauriault, 2014), however, it has become established to interrogate such research-based digital representations and data collections as apolitical spaces (Iliadis and Russo, 2016), to systematically scrutinize power, surveillance and control in the supposed decision support and, if necessary, to help underrepresented interests gain more visibility through counter-data (Dalton and Thatcher, 2014) (Iliadis and Russo, 2016). Dominant perspectives (Kitchin and Lauriault, 2014) manifest themselves - either unconsciously through the organizational structure of the data or consciously as part of a previously developed information architecture of a project - as data and information modelling. In analogy to counter-data, we thus see our approach of what we call concept space analysis as an opportunity to systematically examine data modeling and its social production conditions not only ex-post, but complementarily in the sense of counter-modeling (Kremer/Lang in print) even before their development at the beginning of projects in terms of texamining alternative explanatory approaches.

We illustrate our workflow using the example of a study on place-based narratives about perceived safety in different stakeholder groups in the city of Recife, Brazil (Moura de Souza et al., 2022). A general workflow can be derived (Kremer/Lang in print), which can be explored incrementally in order to obtain complementary answers to research questions:

1 Identify theoretical approach: Which different theoretical approaches can be applied to the research question in principle? Which spatial terms are used to refer to the question from a technical point of view? What are the basic assumptions?

2 Analyze conceptual space: Which spatial concepts should be used to make structures visible on the data in an explorative manner ?

3 Develop appropriate data modelling: In which data schemas should the spatial data be organized? Which explorative data analyses can be applied to appropriately evaluate the validity of the investigated spatial theory approaches with respect to the given research question?