#blogtakeover SWISP Lab x WIL | DATA WALKING | Exploring Data Analysis & Visualisation in School Settings 

A SWISP Lab WIL Project By: Keith Aloysious Soertsz 

My first steps into the inner workings of SWISP Lab were fraught with anxious uncertainty. Speculative wonderings were a new concept for me as I see myself as a pragmatist built on logic and realism. It was only after an illuminating discussion with Dr. Sarah Healy, co-leader of SWISP Lab, that I started to realise the immensity of what they really do.  

WHERE DO I COME IN? 

The SWISP Lab undertakes a number of projects to chart digital stories, social justice issues and imagine futures intertwined with technological metaverses. One of these projects seek to understand education and social issues through data science and analysis. This is where I found myself at SWISP.  

In partnership with SentiOne, Melbourne Data Analytic Platform (MDAP) and Pernix data solutions, SWISP lab collected a large corpus of Twitter data across the COVID-19 pandemic. This dataset spanned every global tweet and news article from 2021-2023 and was catalogued by their influence and sentiment scores.  

Using Pernix’s neo4j Aura database, the corpus was batch-imported and algorithmically topic-modeled using GraphQL and Python programming languages. The refined data was then injected into the neo4j bloom visualisation tool to perform queries and identify relationships between the information nodes. My role was to then establish a line of communication with the data using Cypher queries and computational reasoning.

Two topics were investigated throughout the project:

How did sentiment for climate change fluctuate during COVID-19, and

What were the issues around parenting and hybrid learning during the pandemic?

Although I was only able to scratch the surface of what the neo4j bloom could do with this data, I was proud of my ability to understand data analysis concepts with no prior experience. The project was unsettling and pushed me out of my comfort zone, yet I learned a lot about my own thought patterns and capabilities from it.  

SPECULATIVE DESIGN FOR EDUCATIONAL CONTEXTS 

It is here that I had an epiphany. Schools themselves are deep wells of near infinite data points. What if we were to map and visualise this data in a way that can benefit all stakeholders of a school? What would it look like to use teacher, student and staff data to form a unified network of identities, cultures and backgrounds? How could making these data relationships visible, help form deeper bonds between the human and non-human landscapes?  

I first set out by re-imagining the school environment through a data science lens. I created a representational 3D model of a school setting using Blender and visualised the data-walking cycle through various stages or levels.  

The multi-layered data levels are shown in the video below:

THE DATA-WALKING APPROACH

I then developed a systematic approach to “data-walking” that follows along the pathways of the connective layer. By targeting certain stepping stones, this leads to branching walkways for further data discovery.  The process of this approach is outlined below:

Step 1 – Collecting the data from all stakeholders within the educational environment

At this first step in data analysis, information is data mined by engaging and interacting with all stakeholders in a school. Depending on what data is being gathered, a wealth of information can be extracted about individual or group social, emotional, physical and mental characteristics. From a purely affective perspective, information can be gathered about the latent and hidden aspects of student personality, interests and identities – through such innocuous formats as surveys, questionnaires, journaling etc. The same activities could be asked of teachers and non-academic staff, putting all data points in the same relational field for analysis.

Step 2 – Processing the data through topic-modelling frameworks  

Many advanced-level topic-modelling frameworks are used to breakdown and analyse textual data. In the example above, a student has responded to a data collection prompt about their activities outside of school. What starts as a fairly standard response from a student, can turn into something much more fruitful and abundant. Topic-modelling allows for larger sets of written texts to be broken down into their simplest parts (tokenisation and lemmatisation). This allows algorithms to more efficiently find patterns and relationships in only the data that is pertinent to the study.

Step 3 – Importing the data into databases like neo4j and creating readable visualisations

A diagram of a network

Description automatically generated with medium confidenceThrough visualisation techniques like the one above, relationships between data node points become much more tangible. The relationship model above is a direct translation from the student response seen in Step 2. When combined with responses from other data sources, relationships and groupings start to form within the connective layer. In the example above, a student named Jake has just realised that his geography teacher Mr. Leary also plays badminton on Thursdays and owns a dog too! Jake would have never known this as he rarely talks to Mr. Leary. Perhaps this is a great way for Jake to connect to his teacher and other people in the school community.

Step 4 – Creating visualisation applications to share the data

The last step in the data-walking process would be to share the relational model with everyone in the school setting. This could be achieved through a school-wide open-access application. Students and staff could use the app to find their own relationships and connections with the people, places and things around them. Students could also be taught about data science concepts by using the application itself.

By finding likeness and interests with peers and adults, students can take ownership of their social and affective power in the world. They can use what they learn from the connective layer to form clubs, societies and friend groups across ages, genders and races. An interactive, data-driven model could ultimately lead to a school environment that celebrates inclusiveness, equity, cultural responsiveness and diversity. 

FINAL WORDS

The speculative concepts described in this post are a testament to how deep the rabbit hole goes in terms of education and technology. The 21st century technological landscape, driven by data algorithms, machine learning and artificial intelligence; are reaching new heights in facilitating our understanding of ourselves, and the meta-physical universe. The SWISP Lab therefore aims to launch fringe-dwellers and data-walkers to skirt the line between human and non-human territories – to bring back data artefacts and samples for a/r/tographic study. 

I would like to thank Dr. Sarah Healy and Associate Professor Kate Coleman for their guidance and support in this truly eye-opening endeavour.  This journey has taught me that our reality is far more connected than we realise. In my teaching practice, I hope to harness the power of nodal relationships to strengthen bonds with my future students and colleagues.

I would finally like to thank my WIL Team: Cormac, David, Izzy, Claudia and Bella for their invaluable support and collaboration during this project.

Written by:

Keith Aloysious Soertsz
Master of Teaching Secondary (Visual Arts and Design)
Capstone Topic: ­­Developing Gamified Virtual Learning Management Systems to Enhance Student Engagement in Secondary Schools
September 2024