░░░░░░░ WEEK 7 ░░░░░░░

Graduation Project Physical Presentation

I got some feedback and suggestions from Lasalle Alumnis who have been through Open Studios and Grad Show about the presentation of my work. The points that they brought up from their experience were all things that I was currently not putting much of a focus on.

One point that they brought up was branding, even though my project wasn’t about branding/identity related, it was still important for me to tie everything together visually. They also brought up that during their Grad Show, visitors tended to ignore works that were overly complex and too difficult to make sense of. There are rare cases where some visitors will actually be curious about the work, but that’s mostly driven by visual aesthetics.

Another point they brought up was the presentation of work. Since I had 3 artefacts that were related conceptually but not visually, they emphasised the importance of making it clear that these are 3 tools that possesses the same function. They showed me examples of how their batches did it, as well as directed me to the Lasalle SODC FYP archive.

I found 2 examples on the archive website that I thought were relevant to my project - Urban Plastisphere (Aimee Junoes) and Selfscapes (Rachel Lee). In Aimee’s project, she displayed 3 physical artefacts. Even though they were all made up of plastic, she created a separation by simply placing them on different heights. This consideration allowed me to regard the artefacts as 3 different entities, each communicating a different message but part of the same context/concept. Similarly, Rachel’s project consists of 3 artefacts but it was a mix of physical and digital works. The way she packaged her project together was through 3 books in varying sizes, yet aligning in the same visual style. Each of her artefacts had a name tag along with it so it was easy to refer as well.

Now for my project, I’ll likely have physical outcomes to go with my digital ones. The physical formats are to document the success of the artefacts in provoking reflection, as well as the process of the experiments. Even though everything falls under the same context, I wanted each individual artefact to have their own unique visual style since each of them adopted a different approach (Definitely also to showcase my flexibility).

Since I’ve decided on individual stylings, I would then need to find a way to package them together under 1 concept. When I was discussing with my friends, I proposed that I should come up with 1 small brochure for each artefact, with the same styling. This idea was quickly shut down due as I would have too many printed collaterals and it would just be an overload of information. It would be much better to design sections, or just 1 printed pamphlet to tie everything together. I drew up some quick wireframes of what I’m referring to based on last year’s individual table layouts. These are my options for now, but when we get closer to the open studios date, I will lock everything down.

Urban Plastisphere - Aimee Junoes


Selfscapes - Rachel Lee


Setup Wireframe A (Sections)


Setup Wireframe B (Pamphlets)

Artefact B Conceptualisation Process

The primary objective of this artifact is to prompt individuals to contemplate the dilemmas of algorithm transparency, which entails ensuring that algorithmic processes and associated data remain accessible to the public. Moreover, I aim to achieve this goal computationally by employing a sequence of visuals crafted in TouchDesigner. A persistent challenge I encounter in computational design pertains to the utilization of imagery. Upon examining computational outcomes, I observed predominantly three types of visual elements employed: images, typography, or renders. Despite experimenting with renders during the initial semester, I found it challenging to integrate them cohesively within the context of my project. Similarly, typography posed difficulties in being effectively contextualized to my advantage. Consequently, I made the decision to prioritize images as the primary element in this artefact.

While developing my initial artifact, I stumbled upon the concept of live folders. These folders enable the visuals in my TouchDesigner file to adapt to live content updates, a feature I found particularly intriguing and wished to incorporate in some capacity. Given my focus on algorithmic curation, I recognized that social media images would be the most relevant context. Interestingly, I often find myself prioritizing technical considerations over ideation. This time, however, it proved advantageous. While exploring methods to download images from my Instagram explore page, I delved into experimenting with the Chrome inspect tool and encountered a galore of cached images.

The alignment between using cached images and the concept of live folders proved to be quite fitting. Cached folders retain every image encountered following a specific action (such as scrolling, clicking, or refreshing), making them an excellent fit for an interactive component in my project. However, I encountered a significant obstacle. Legally, accessing Google Chrome's cache folders is not permissible. They are stored on a cloud network, making them inaccessible for download. Nonetheless, the browser does permit the individual downloading of images, which is the approach I opted for.

Up to this point, I haven't finalized my decision on what I'm going to do. I've simply collected all the resources and tools necessary to construct a narrative. Since the narrative will heavily rely on my visuals, I'm comfortable with exploring until I gather enough resources to formulate one. While browsing Reddit, I stumbled upon a post where an individual unintentionally produced a visual outcome while aiming for something entirely different, and his friends dubbed it 'the creative process.' This resonates deeply with my experiences in computational design, I guess it’s really a universal experience.

I do have one experiment I want to revisit. In the first semester, I conducted a visual exploration of the 'algorithm black box,' where the inner workings of algorithms are concealed. This encompasses the decisions they make and the data they utilize, among other factors. If I can find a way to make it provocative and encourage people to reflect on algorithm transparency, it will undoubtedly be for the best.

Cache Folder Inspect Tool


'The Creative Process'

Noise Displacement & BlobTracking

As I was looking for TouchDesigner tutorials regarding image manipulation, I came across the concept of BlobTracking. Basically it’s similar to the Inclusivision software’s face detection, but instead of detecting faces, it detects movement and gives you the x and y values. With these variable x and y values, we can draw materials/text over them to mimic the movement.

In the tutorial, they used a noise map to displace the input video and then tracked the displacement. This made me understand 2 things. Firstly, I can use a noise map to create random algorithmic distortions. Second, any sort of movement can be tracked, so I don’t exactly have to follow the same approach.

After trying the blobtracking tool out myself, I realised that there was a limitation with the basic subscription of TouchDesigner. In the tutorial video, there were a lot of points being tracked and the creator was able to connect all the lines to form this web. However, for my version the maximum number of blobs I can track is 2, meaning that I’ll only have 1 line connecting the 2 tracked points every frame. This might seem minor but it resulted in a huge difference in the impact of the visuals. The one in the video seemed overwhelming (but controllable), but mine was plain boring. However, in my version, you could see the tracking of the movement clearer as there was less to look at.

Thinking of how I can apply this to my project, instead of using 1 image, I could displace and track multiple images so the lines could be drawing some sort of connection between the images. Using cache images from my own Instagram explore page, I replicated them floating in random x y z positions and tested it out. The result wasn’t what I had expected, it was boring, in fact more boring than when it was just 1 image.

Having not achieved my desired outcome initially, I attempted another method. This time, I created a noise map and had the blobtrack TOP track its values. Instead of directly applying the noise map on my images, I utilized its channels, allowing for more lines to be drawn without being limited to just one. I didn't quite appreciate the fluid effect demonstrated in the tutorial. Given our focus on algorithms and data, the fluid motion didn't quite reflect the movement one would anticipate from algorithms. To synchronize the images with the movement of the lines, I applied a glitch effect I had previously learned on top of the images, creating the illusion of motion. This also gave the impression that the images were being scanned, which ultimately worked to my advantage.

Despite using the same tools, the version of the blobtracking with multiple images and the version with a single image both has its own unique expression. The lines in the one with multiple images look like it was trying to draw connections between images, and the lines in the single image looks like it’s analysing the pixels within the image. I think this was the point I was starting to realize my visuals were veering toward interpretations of the algorithmic curation process. I wouldn’t mind that as long as this creative churn continues, but I was also quite worried about the provocative portion of my project, since it is still absent right now.

Limited to 2 Blobs


Multi-Image Tracking w/ Lines


Glitch Effect

Designing Provocation

As I continued my search through TouchDesigner outcomes on Pinterest and Instagram, I sought inspiration that could infuse a sense of provocation into my project. Despite stumbling upon some intriguing image manipulation visuals, none seemed to align with the current concept of 'Interpretations of the Algorithmic Curation Process.'

With my concept already in place, I shifted from the 'accidental discovery' creative approach to delving deeper into the relationship between cache data and algorithmic curation. Initially, I assumed my understanding of cache was comprehensive, only to realize its limitations upon further research. While investigating, I discovered that there was no direct link between how Google Chrome stores Instagram Explore Page images and the cache utilized by Instagram for curation algorithms, which resides on its independent cloud server. Nevertheless, I realized I could frame my narrative of Chrome's cache data as Instagram's cache data.

As I pieced together the puzzle, I began to question whether my identity was a subject of curation. While social media algorithms learn from our digital patterns and behaviors, there comes a point where individuals encounter new and different discoveries recommended by these algorithms. At this juncture, do our identities become subjects of algorithmic curation, or do our identities continue to inform these curation algorithms? This revelation prompted me to contemplate further how these algorithms function and what kind of data they extract from us to recommend these fresh discoveries. Consequently, I intend to frame my interpretations of the algorithmic curation process as a narrative that prompts viewers to ponder this question.

Visual Narrative

I’ve always liked to include an interactive element in my computational work, so for Artefact B, I would like to do the same. Preferably, I would like to create a visual installation and get some friends to be my subjects, offering me their Instagram Explore page cache images. I don’t think anyone I’m not close to will be willing to let me invade their privacy like that. Using their cache images, I would then create a visual interpretation of the how these images will be analysed, sorted in the algorithm black box, and then becoming a part of their identity. To do this, I broke down the installation into 5 phases:

Phase 1: Collecting Cache Data (No Visuals)
Despite my desire to streamline this process, legal constraints prevent me from downloading the images as a folder, so I’ll have to download them one-by-one. I truly wish I could access the live folders to create a more immersive interaction between the visuals and my subjects. Just imagine if they could scroll their Instagram explore page and witness how these visuals change and update in real-time. Unfortunately, I must work within my current limitations. In order to avoid introducing human bias into the data manipulation process, I will select these images at random without previewing them. Additionally, I'll be mindful to exclude cache folders containing profile icons to only use posts.

Phase 2: Uploading Cache to the Cloud (Visuals)
I want to create a visual interpretation of the process behind uploading the images to the curation database. With my improved proficiency in TouchDesigner, I anticipate being able to plan ahead before creating, rather than relying solely on accidental discoveries. However, I must note that accidental discoveries can yield excellent results, and if they surpass my initial vision, I fully embrace them. During this phase, I also intend to incorporate the cache image names. I've observed that the cache images I downloaded possess unique and systematic names. Leveraging this data can assist viewers in comprehending the process unfolding before them.

Phase 3: Cache Database (Visuals)
The objective of this phase is to illustrate the vastness of the data stored in the cloud. I plan to arrange all the images and manipulate the z-axis to craft an endless landscape. Through this phase, I aim to evoke a sense of the overwhelming scale of the database. Despite only utilizing a selected number of images, I hope to convey to my subjects the magnitude of their data—from the inception of their Instagram account until the present—coexisting within the same database.

Phase 4: Cache Data Analysis (Visuals)
In this phase, I aim to leverage the line visuals from the multi-image blobtracking, where the lines appear to be forging connections between images. These visuals are open to interpretation and serve as a transition to the next phase. Rather than flatly presenting the images as in phase 3, I intend to introduce a different type of movement to imply a sense of depth and dimensionality within a 3D space.

Phase 5: Identity Consumed by Cache Data (Webcam + Visuals)
In this final phase, the narrative comes full circle as I aim to illustrate how cache data is gradually assimilating our identities. The primary objective is to pose the provocative question mentioned earlier, prompting viewers to reflect on its implications. To enhance immersion, I plan to integrate the webcam into this phase, deepening the viewer's engagement with the experience. Drawing from my previous experiments with the ASCII filter, I intend to explore the feasibility of replacing symbols with images instead. The visual I aspire to achieve involves substituting all the pixels in the webcam's input feed with the cache images, creating a compelling visual representation of the narrative's climax.

Phase 3 Reference


Phase 4 Drafts


Phase 5 Reference