Hello! I am Xinyu, a graduate student
at University of Leeds
MA Digital Media Program.

Hello, I am Xinyu, a graduate student at University of Leeds. I was born in Jinhua, Zhejiang Province, a small city lying on the south-east coast of China, famous for ham and bergamot. After graduating high school, I went to university in Shanghai and lived there for five years.

*Just a friendly reminder for Holly, my weekly reflections on the Workshop are at the bottom of the webpage :)

• Learn More

Recent Work

Guangzhou and "Globalization from below"

Images and Visual Ethnography on Guangzhou, a Metropolis in Southern China.

Ultricies lacinia interdum

Lorem ipsum dolor sit amet nisl sed nullam feugiat.

Tortor metus commodo

Lorem ipsum dolor sit amet nisl sed nullam feugiat.

Quam neque phasellus

Lorem ipsum dolor sit amet nisl sed nullam feugiat.

Nunc enim commodo aliquet

Lorem ipsum dolor sit amet nisl sed nullam feugiat.

Risus ornare lacinia

Lorem ipsum dolor sit amet nisl sed nullam feugiat.

• Full Portfolio

Workshop Reflections

(Semester 1)

Week 1: Overall, it was pretty good, but I felt like we spent too much time registering an account and connecting to the FileZilla.

Week 2: It felt like the classmates hadn’t done the necessary preparation before class, which slowed down the pace of the lesson.

Week 3: The explanation of web scraping was more introductory. It would be more beneficial for our future research and work if we could actually use an easy-to-handle scraping software to gather data and engage in some hands-on activities.

Week 4: This week's group project was great; it felt like I was back in the groove of doing academic research.

First half semester: The the biggest thing I learned from the first half semester was building my own website and overcoming my fear of coding. The most challenging thing was to learn unfamiliar computer software by myself and apply it to the assignments. Data visualisation is powerful because it allows the data to form a sort of ‘narrative’, a powerful way of making sense of the world. For example, data journalism can help audiences understand facts in a more innovative way just by presenting visualised data. However, there are some limitations to the visualisation of data. For example, errors and subjectivity in data clarity and selection, which can lead to biased results in digital narratives.

Week 5: Data visualization enables the presentation of information in a concise and logical manner, thereby constructing a form of "narrative." However, it also faces challenges such as the subjectivity involved in data selection.

Week 6: ...(to be continued)

Week 7: Through training my own model, I have discovered that machine learning is an incredibly powerful tool. By analyzing vast datasets, it allows us to develop a wide range of applications, such as facial recognition systems and emotion detection systems. However, the training process is not without its challenges. For instance, AI tends to perform less effectively in facial expression recognition compared to gesture recognition. Furthermore, malicious input of incorrect data can undermine the accuracy of AI models, and the potential for racial and gender biases in machine learning datasets poses significant ethical concerns. This workshop also illuminated the possibility of machine learning becoming a component of Foucault's concept of the "panopticon," raising concerns about its potential role in surveillance and control.

Week 8: I think ‘social world’ means the social network we are in. For us in the Internet age, ‘social world’ is twofold: the social world of physical space and the social world of cyberspace. Digital ethnography expands on traditional anthropological methods by placing the occasion of fieldwork on the Internet. The analysis of Internet images, text and sound, in-depth interviews and focus groups by means of video calls, etc., exemplify the flexibility of ethnography in practice. I believe the most important thing for the ethics of digital ethnography is to protect the privacy of the research subjects. We cannot publish their data on the Internet in academic journals without their permission.

Week 9: To improve Sumpter's methods, I would use web scraping tools to collect social media data, reducing repetitive manual work in research. I’m also interested in studying my friends’ main and alternate social media accounts—how they switch between them, the differences in the content they post, and how these reflect the fluidity of their identity. I’ve realized that digital communities, much like physical ones, can form a collective memory for their members. I’d like to explore how digital communities shape group identities and examine the differences between how collective memory is formed in digital versus physical communities.

(Semester 2)

Week 14: Through this Special Collections Field Trip, I learned to extract valuable data from historical documents and draw meaningful conclusions. For example, we found that in 19th-century archives, students' ages varied widely, with some starting school as young as four. This suggests that church schools then were more like nurseries than formal institutions. A major challenge was deciphering the handwriting, but AI tools like ChatGPT helped with text recognition. I think selecting documents with more substantial information would improve the project, as many we examined lacked enough data for effective analysis.

Week 15: Datafication can be defined as the transformation of human life into data through processes of quantification, and the generation of different kinds of value from data. (Mejias and Couldry 2019) I will select examples from the literature showing how data collection was historically controlled by those in economically privileged positions. Even today, data remains concentrated in the hands of governments and tech giants, showing that data control has always been a privilege. The key difference is that modern control is more hidden and pervasive. For example, our online activity is stored as code, making it harder for those without technical knowledge to access. This has widened the gap in data accessibility, making inequality even more severe than in the past. Through reading the LGBTQ+ Collection Guide, I realized that data is not neutral—it carries power and bias. For example, when LGBTQ+ individuals were not socially recognized in the past, they might have concealed certain details when recording their data. As a result, the data itself reflects inherent biases. As a result, when collecting and analyzing data, we must recognize its embedded ideologies and approach it critically to ensure a more informed and balanced interpretation.

Week 16: "History is about the production of evidence-based arguments, not about telling the truth." He believes history is not about finding absolute truth but about making reasoned arguments based on evidence. Historians interpret sources, and their narratives are shaped by perspective, not just facts. It means digital historical sources are made up of multiple layers—hardware, software, metadata, and networks. These layers affect how data is stored, accessed, and interpreted. Digital sources are not fixed like paper documents; they change depending on technology. When digitizing historical materials, several issues must be considered. First, digital copies may lose physical details and context. Moreover, they are not exact originals but new versions. In addition, file formats affect long-term access, and metadata is essential for tracking authenticity. Furthermore, digital files can be altered, raising concerns about data integrity. However, while digitization improves access, it may also create a false sense of authenticity. Therefore, careful planning is needed to ensure reliability.

Reflection on Week 15: The Week 15 workshop is interesting! Through group discussions and exchanges, we reconstructed a historical detail. It's a fascinating experience in historical data research!

Week 17: According to this week's literature, archives shape knowledge by selecting, categorizing, and structuring data. They determine what is preserved, how it is accessed, and how it is interpreted, influencing public understanding and decision-making. Archives often exclude marginalized voices, alternative perspectives, and inconvenient data. Economic and political interests, as well as biases in data collection, contribute to these omissions. Uncertainty refers to the biases, gaps, and unpredictability in big data archives. It challenges the reliability of data while also being used as a tool for control, governance, and commercial exploitation. I think my data collection may have some biases. For example, the identities and narratives of LGBTQ+ individuals are simply classified into a binary gender framework. I understand that I should use more information technology tools to simplify repetitive tasks, such as OCR.

Reflection on Week 16: To be honest, I don’t quite understand the visualization part of the group work. I feel like this week's workshop was not presented in a logical structure.

Week 18: Feminist data visualization challenges biases in data by questioning traditional categories, amplifying marginalized voices, and emphasizing context. It ensures that visual storytelling is more inclusive, making hidden inequalities visible rather than reinforcing dominant narratives. Symbols and colors can bridge past and present by visually linking historical themes to modern issues, while captions provide context and guide interpretation. Editorial choices shape how these elements are presented, creating a compelling and meaningful narrative that resonates with today’s audience.

Reflection on Week 17: To be honest, I don’t have a clear understanding of the task we need to complete this week. Compared to last semester, the assignments this term feel less specific and harder to execute. Our group still hasn’t decided what "new data" to collect because we first need to revisit the University of Leeds archives for 19th-century child labor data before determining our new dataset. During the first field trip workshop this semester, we didn’t realize that the photos we took would be used for weeks of in-depth research, leading to insufficient data collection. Without a solid foundation, finding "new data" and creating visualizations has become difficult, forcing us to start over. The biggest challenge is that the workshop tasks this semester feel too abstract, lacking concrete instructions to guide our work. For example, we are simply told to collect "new data" from the internet without any clear criteria. This kind of uncertainty is something I haven’t encountered in other modules. From my point, clearer guidelines on assignment formats, such as the expected form of "new data" or the type of group video submission (e.g., PowerPoint recording or another format), would be helpful to guide our work.

Week 19: Digital media engage multiple senses by combining visuals, sound, touch, and movement. Technologies like video, VR, AR, spatial audio, and haptic feedback create immersive, interactive experiences beyond traditional media. It can also enhance research by enabling rich sensory data collection, interactive participation, and immersive representation. Tools like video, wearable sensors, GIS, and VR help capture, analyze, and recreate sensory experiences more effectively. In our group video, we will use visual and auditory sensory experiences. For example, we will play a documentary about child labor in 19th-century factories to help the audience immerse themselves in the historical context of our research subjects.

Reflection on Week 18: We haven’t started visualising data yet because our group needs to reschedule a visit to the Leeds University Archive to collect more historical data on 19th-century child labor. After that, we’ll search for new data online before moving on to visualisation—I think most groups are at a this stage. In the first workshop this semester, we didn’t realise the field trip archive was essential for later workshops, so we took few photos. The biggest challenge is still that there are no clear guidelines in our workshop tasks. What kind of new data should we find online? What software should we use for visualisation? What format should our group video take? Are there any examples to follow? We really some examples to help us understand what to do. XX

Reflection on Week 19: The Map Sensory Experiences workshop was truly engaging. It made me realise that digital media isn’t just conveyed through sight and sound but can also be experienced through taste, touch, and other senses. The biggest challenge lies in how to visualise and integrate sensory data effectively. In our group project, we could experiment with this approach, leveraging multiple senses to create a more immersive experience for our audience.