Week 6: Project Proposal 1 / by Valzorra


While I was researching Markov Chains and Matrices, I explored the concept of On-Page SEO, specifically what variables contribute to websites ranking higher in search engines. The conclusion of my analysis was that one can give numerical values to the key On-Page SEO Factors, and then estimate what changes need to be implemented to achieve the ranking one wants. In simpler terms, we can figure out what to change to rank higher based on where our site is at and where we want it to be (for the full analysis, please refer to Week 4: Probability Manipulation in SEO). However, although I have explored the relationships between the variables, their relative weights, and how they could fit into a Markov Chain, this process can get quite technical. Many website owners and publishers are not necessarily familiar with Matrices and Markov Chains, so the knowledge of how to improve their performance would require unnecessary training in a field that’s irrelevant to their work. However, I believe I may have found an answer to this problem through a new methodology for Data Visualisation (which is closely related to my research on the topic, found in Week 5: Mathematics Visualised). What’s more is that this method does not necessarily need to be limited to SEO, but it could rather be a universal visual training tool, which would mitigate the issue of working with raw numbers.

Project Overview

Most methods for Data Visualisation take place in two-dimensional space, such as histograms, point graphs, arc diagrams, and more. These techniques are quite efficient when working with a very limited number of variables and when the relationships between them are straightforward. However, it would be rather difficult to depict more complex and intertwining information. For example, On-Page SEO Factors roughly fit into three major categories, with a series of three to five subcategories, all of which affect each other differently. In order to represent this complex information, we need to take it into an additional dimension. By representing the data in three dimensions, we gain access to a series of solids, with a variety of faces, edges, angles, and vertices, all of which can be used to represent data. For On-Page SEO Factors, I have found that a tetrahedral shape represents the variables in the most clear and concise way possible (for more on tetrahedrons, please refer to Week 3: Research on Geometry). The three major categories of On-Page SEO would be represented by the base of the tetrahedron, while the pinnacle of of the solid would represent time.

Inputting different data points into this model would then change the tetrahedral shape into a different triangular polyhedron. A perfect tetrahedron would represent the ideal state of a website in terms of search engine success, while any other polyhedron that differs from the Platonic Solid would show what areas need to be improved. By inputting data from different websites into this model, we would essentially create a library of polyhedral shapes. That way if a website owner wants to improve their ranking in search engines, they wouldn’t need to learn a thing about Matrices or Markov Chains. All they would need to do is ensure that the shape of their website either matches the shape of a website they are aspiring to be like or is as close to a perfect tetrahedron as possible. What’s more is that this could also be a fantastic tool for education and training. For example, it could be presented to Marketing students, who would be tasked with creating a successful marketing strategy based on certain variables. The strategies the students have come up with, would then be inputted into the tool and all of the generated shapes would then be compared to each other and to the ideal. This would provide a clear and visual method for training and education, and could be applied to any field that uses some form of variables.

What’s even more exciting is the possibility to make this project a real-time interactive tool through the use of VR. Data Visualisation in VR has not been explored greatly and most efforts thus far have not taken great advantage of the medium. However, I believe this tool would be perfect for VR, because of its immersion and incredible interactive potential. Instead of inputting the data points for this model through a keyboard, one would be able to physically move them and change the data in the Virtual Reality Environment. As one grabs a data point and moves it about, changing its values, the entire shape would change then and there. Placing the model in VR would make the process a lot smoother and faster because moving the data points physically is extremely intuitive and can be much easier than typing different values in. At the end of the experience, the tool would provide concrete numerical data of what needs to be done to achieve the desired shape, allowing users to directly implement the suggested changes.

Concept Pieces

The concept piece below roughly shows how the model may look in the 3D environment. All the different coloured points represent data that could be shifted about to form a different shape, and how the connections between them would change. The larger concept piece represents a top-down view of the model, while the smaller one to the side shows what one of the faces would look like to the side, with its respective data points. The second concept image explains the process of taking the graph into the third dimension and constructing the shapes off of that, thus forming an elaborate library of shapes.


Thoughts and Reflection

Overall, I am very satisfied and excited about this model because of its universality, usefulness and applicability. This is a method for representing any sort of data and can be used as a visual training tool for a multitude of fields. Additionally, it is a great way for the communication of complex data and its intertwining relationships. The tool can be developed and used on computers, tablets, phones, and it could even be a potential breakthrough when it comes to Data Visualisation in VR. The only problem with it so far is that this is very much focused on data and is not really related to games, which can make the model a bit dry for an FMP. Nonetheless, I do believe this is a project worth further research and pursing because of its incredible educational potential, and its value as a piece of clear and concise piece of design.