Digital inequalities, placemaking and geographical diversity in tweets about Cape Town railways, 2010 - 2022
DOI:
https://doi.org/10.55492/dhasa.v7i02.7703Keywords:
Geographical bias, digital inequality, Twitter, digital placemaking, LLMsAbstract
Digital placemaking is implicated with power relations and can expose inequities, exclusions and erasures. A longitudinal archive of 5606 tweets geotagged in Cape Town was collected to explore placemaking practices and inequality in commuter posts about rail travel in the city over the period 2010--2022. Named Entity Recognition (NER) identified 1755 tweets from this archive as a record of digital placemaking about the major train lines in Cape Town. A sustained bias over time in favour of the suburban Southern Line is detected. Strong core-periphery patterns emphasize major interchanges in the collection, while measures of diversity indicate small increases in geographical diversity in the railway system over time.
Structural biases in access to the means of digital placemaking thus shape this archive. While not causally related, a similar geographical skew towards socially hegemonic places in the system was apparent while using an LLM to annotate a sample of the tweets (n=350). Algorithmic biases were even more pronounced in a synthetic sample of LLM-generated tweets about the railway system, which overwhelmingly focused on local attractions on the Southern Line. Thus the Generative AI tools used to read the archive themselves introduce algorithmic biases and potentially perpetuate socially hegemonic digital defaults.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Marion Walton

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.