RPO Feedback Breakdown
Refine Terminology
A major takeaway was the need to refine my terminology. Right now, I've been using terms like data double, digital twin, and algorithmic identity somewhat interchangably. The feedback pointed to the importance of clearly distinguishing them.
⑴ Data Double
The fragmented digital version of a person assembled from data
traces which is essentially a “ghost self” produced by surveillance systems (Lyon; Haggerty and
Ericson). These doubles exist independently from the individual and are used in predictive
profiling and decision-making (Cheney-Lippold).
⑵ Digital Twin
A real-time, dynamic simulation of a physical system or
person, built for monitoring and prediction (Rasheed et al.). Unlike the data double, it
maintains an active, bidirectional connection with its physical counterpart, often used in
engineering, healthcare, and smart systems (Emmert-Streib et al.)
⑶ Algorithmic Identity
Describes how systems infer and classify individuals
based on data patterns, creating operational identities shaped by algorithmic logic
(Cheney-Lippold). These profiles affect how we are seen and treated by digital platforms,
institutions, and systems (Kotliar; Onitiu).
RPO Reframing
Through ongoing readings, experiments, and consultations, I refined the conceptual focus of Pillar 1 and Pillar 3 in my project. These reframings clarified how the work moves from understanding how identity is constructed by algorithms to exploring how these processes can be communicated, felt, and reinterpreted through digital exhibition design.
Pillar 1: Datafication and the Construction of the Data Double
After working through multiple readings about Datafication, I realised that "algorithmic identity" is the most suitable term for my research. Unlike concepts that simply describe data collection or representation, Cheney-Lippold’s framework captures the full process of how identities are constructed, interpreted, and acted upon by computational systems.
These identities continually shift as new data is gathered, revealing how deeply human behaviour becomes entangled with systems of prediction and classification. Yet this process remains largely invisible, surfacing only through personalised feeds, recommendations, and targeted content, setting the stage for understanding how algorithmic feedback loops shape and manipulate our digital experiences.
Pillar 3: Digital Exhibition Design
Lev Manovich’s work demonstrates how digital exhibition design can reveal the hidden structures behind cultural data. Through large-scale data visualisations and computational analysis, he exposes how algorithms classify, pattern, and organise culture at scale.
His practice shows how abstract processes such as cultural analytics, sorting, and pattern recognition, can be translated into visual forms that reveal the logics shaping digital life. In the context of this research, Manovich’s approach underscores how design can make algorithmic operations visible and interpretable.
Refik Anadol expands digital exhibition design into immersive, sensory environments that visualise the inner workings of machine learning systems. Rather than simply analysing data, he transforms the computational process itself into a living, aesthetic experience, bringing audiences inside the “hallucinations” of AI.
Through monumental, data-driven installations, Anadol demonstrates how algorithms can be communicated emotionally as well as intellectually. His work illustrates how digital exhibition design can turn invisible systems of data, memory, and prediction into tangible encounters that evoke awe, reflection, and critical awareness.
Prototype 1
For Prototype 1: Feed Your Garden, I wanted to explore my readings on data collection and the mechanics of personalised social media feeds. These systems learn from every gesture: scrolls, pauses, clicks, and use them to cultivate a feed that feels uniquely “ours.” Using that as the grounding metaphor, I wondered: if our actions feed the system, what happens when we feed it with our words instead?
This prototype tests whether language, not behaviour, can become the input that shapes the system’s response, mirroring how platforms continuously construct and refine our algorithmic identities.
Read more about the algorithmic process breakdown in Prototype 1's Catalogue of Making.
Try Prototype 1 Catalogue of Making (Prototype 1)
I want to explore how algorithms interpret the data we feed them, shaping how our identities appear within digital systems. I wanted to visualise this invisible process: how traits, preferences, and behaviours are translated into algorithmic profiles that grow and evolve over time. Using the metaphor of a garden, the experiment invites users to "feed" traits they associate with themselves, which in turn generate a unique digital ecosystem.
Each trait functions like a data point: a seed, that influences how the garden develops. The system processes these inputs through simple algorithmic rules that mimic classification, weighting, and prediction. Over time, patterns emerge: while every user’s garden grows differently, shared traits create moments of resemblance, revealing how individuality and sameness coexist within algorithmic systems.
p5.js
I gave the system different traits as a pre-defined dictionary, and with the help of ChatGPT, I asked it to draw many different floral shapes in order to allow each shape to correspond to a trait. Just as every flower species has its own form and growth pattern, each trait holds distinct characteristics that shape the overall composition of the garden. These floral variations differ in shape, colour, movement, to visualise how diverse inputs coexist and interact.
I assigned specific words to different floral shapes to translate quanlitative traits into visual forms. For example, words like "gym", "workout", "dance", each generate distinct flower types that I assign it to. This allows for personality and emotion to take shape through the visual variation.
The act of "planting word seeds" reinterprets the way we feed algorithms through our daily interactions: likes, shares, comments, etc. Instead of the behavioural data, I want my users to be able to input traits or interests they associate with themselves, such as "funny", "weird", "loud", or things they are interested in like "gym", "workout", "pop".
Each word becomes a seed that once planted, generates a visual element within their digital garden. This shifts the usual algorithmic exchange into something more intentional and self-reflective. Users consciously choose what they want to feed into the system, watch their data grow into a living representation of their own traits and identity.
Challenges Faced: The dataset I used is quite small. Only a limited set of words were mapped to forms. this contraint reduced the diversity of the visuals, making the garden less dynamic and representative of nuanced traits.
What Could Be Done Better/For Future Expansion:
• Expanding the word database
and floral shapes and refining the mapping rules can allow for richer, more complex growth
patterns that better mirror human individuality.
• Include interactive toggles or sliders
that let users decide how each trait is represented, give
them agency over how their identity is visualised and make the experience more personalised.
After users finish planting and cultivating their gardens, I added the "Save" button so they can save their creation and enter "View Gallery". This "View Gallery" stores all the saved gardens, each representing different users' traits that are collectively displayed. This shared archive extends the idea of algorithmic identity beyond the individual.
Just as platforms aggregate and compare user data to form patterns and profiles, the gallery
visualises how individual inputs coexist within a larger ecosystem. It reveals both diversity
and convergence: how personal data, when placed side by side, begines to form visible trends,
overlaps, and algorithmic groupings.
By including this gallery, I wanted to reflect how
our seemingly unique identities online are constantly archived, compared, and shaped in relation
to others within the algorithmic system.
Visual Archive
