data ethics data feminism eg feminist data set by Caroline Sinders https://carolinesinders.com/feminist-data-set/ STEM work by and for underrepresented minorities data science as a newer form of power empirically wrong data categorisation, incorrect and unjust binaries data visualisation datasets that are skewed, incomplete or lack diverse representation medical studies historically exclude black women data collection on trans people very poor i’ve seen a bunch of papers on this i just haven’t got round to cataloguing them here data witnessing Mentions “All data is historical data, the product of a time, place, political, economic, technical and social climate. If you are not considering why your data exists and other datasets don’t, you are doing data science wrong.” — Melissa Terras (quoted in How AI Image Generators Make Bias Worse) ...
data visualisation communicating with non-data people data doesn’t belong to the specialists starts in visual expression through glyphs, pictograms etc in The Age of Noise we have so much data we need more data visualisation more and more now as we have more and more data human attention spans decreasing Why is data visualisation effective? Theories from cognitive psychology, design, gestalt processes, neuro-aesthetics, science of decision-making and persuasion Sight most cognitively powerful sense the speed at which we process visual information is roughly equivalent to the bandwidth of a computer network (?) we disregard much of the information we are processing about 0.7% of what we perceive is what we are aware of RGB is based on our own anatomy (what ocur rods and cones process) but CYMK on ink chemistry reflected light turned into electrochemical signals, through optic nerve into occipital lobe in the back of brain all senses have two modes of operation, pre-attentive processing (system 1) and attentive processing (system 2) System 1 is quick/automatic/habit/subconscious/irrational/impulsive and System 2 does long-term planning, conscious thought, calculation, rational, deliberate, more precise system 2 more ‘trustworthy’ I agree for the most part but system 1 is wired for survival, so trust those trauma instincts if they still serve you, they’re there to protect you consciously processing stimuli must be done in serial which is why it is slower (I swear parallel thought is a thing though? Or maybe just very fast serial) when making data visualisations you’ll be in system 2 but a viewer is likely in system 1 (or at least that’s what they use first) use visual characteristics deliberately to trigger instinct in the viewer, eg colour, form, shape, size, distances, spatial presentation, motion (fast-twitch in particular) gestalt principles ie proximity, enclosure the whole is greater than the sum of it’s parts it takes effort to even start switching to system 2 so we often stay in system 1 Resources https://www.youtube.com/watch?v=KPdQykj9a04&t=380s ...
data extractivism eg as applies to less-than-ethical research with eg indigenous communities but also as applies to users of Web 2.0 services and platforms, digital forms of data, hidden data flows and data exhaust, the ‘extraction imperative’ that sees wasted opportunity in possibilities for data collection to be sold on to train predictive models (AI ethics) users forcibly becoming data subjects more data is extracted than what is needed for the service beyond online life - data from daily experience, mics in products serving the business model and corporate interest ...