Week 2

Writing web pages using html is a very new experience for me and it always makes me jump at the chance to try it. But right after that, it can be just as frustrating for me. First of all, it is very strict on spelling and formatting, so my carelessness always made me suffer a lot, I would spend a lot of time checking, even sometimes I couldn't find out where the mistakes were even after checking many times. In addition, for an international student, the meaning of href/rel etc. codes is very confusing to me because I don't know what they mean, so I have to go back to the information over and over again in the process of writing, which also shows that my proficiency in the use of these codes still needs to be improved. Thirdly, when using the learning materials on the site, I was also lost in the jargon, and in many cases I couldn't even understand exactly what the creator of the tutorial was trying to say.

Week 3

During this week I have started to try and build my own website and I have realised that there are a few things that I hadn't noticed before that are essential to writing html such as etc. I have also come across a lot of unexpected things when I actually start to build the site, even though I was confident in my foundation before I started. Such as the fact that css can only be used in html if the html and css files are in the same folder, and the fact that web pages crash when you change the location where the html and css files are stored on the way to building them. Additionally I can't accurately estimate the size of the image on the page and when I change the value it creates a problem I didn't expect, the image isn't scaled up proportionally, it tends to change the aspect ratio due to the change in value. These problems were just one of the problems I encountered, there were many, many more that were totally unexpected and I had to spend a lot of time searching for solutions (even some that seemed to be unsolvable) because what I learnt in the previous tutorials was rather basic and they didn't solve my problems well.

Week 4

This week I tried to use some templates to help me write my website, when I downloaded the templates and unzipped them, I was a bit overwhelmed by the meaning of the many files in them, and after many attempts I was able to find the html script and open it as I had hoped. In the subsequent modification process, I often could not find this title or text corresponding to the script, as well as some of the new code appeared to be very confused, unable to make accurate judgement as well as modification, resulting in my redo a number of times. In addition, when I want to add something other than a template to a web page, they are always very difficult, I need to write my own “style” about it, such as the size, position, etc., which I need to consider, and even sometimes when I try to do it myself, it will cause the whole web page to crash, and the web page will become blank in the preview, and the colour of the text in the HTML text will be The colour of the text in the HTML text would change (I still don't understand what went wrong), and I would have to undo it again and again to recover my page.

Week 5

This week we have been doing some hands-on data related activities. Firstly, after some discussion in class, we chose a form of data collection that was defined as ‘researcher-led’, which at the same time allowed us to learn about the diversity of forms of data collection. But honestly, after some discussion, our group regretted this choice because this form of data collection did not allow us to find suitable cases so easily. But in the end, with our continuous attempts, we determined our research direction, virtual chat partner. This is actually a kind of conversational AI, based on artificial intelligence technology, which interacts and communicates with users by simulating human conversational behaviours to provide emotional support, companionship or entertainment. And we chose the software Replika as our research object.

Week 6

In order to collect the data we wanted, we went through the Apple store for relevant reviews about the software Replika, and for the sake of diversity of data and equally to prevent the problem of high overlap in the data collected, I proposed that we divide the work, i.e., each member of the group collects according to how high or low he/she rated the app. I chose to go with low rated reviews. During the collection process I soon noticed something, among these low rated reviews I found some reviews complimenting the app, which was perhaps sarcasm from the users? Maybe it was a special case, but I couldn't confirm it, so I chose to discard these comments for the sake of data reliability. In addition, there were also a lot of unexplained bad comments in the low-scoring comments, i.e. the comments posted seemed to have nothing to do with the software itself. ...... In the process of data visualisation since then, through practice, I think a good data visualisation needs to be a good one, which is a good idea. A good data visualisation needs to be concise and clear, e.g. the focus needs to be on the key points, etc. to ensure that it is readable. This reminds me of the ‘kiss’ principle that my tutor suggested in my previous undergraduate studies, i.e. keep it stupid simple, and the principle of interactivity in data visualisation should also be considered.

Week 7

In this week's Workshop we trained the Image Recognition AI, which I have to say was a very interesting experiment, and judging by the results it's incredible to me. I also found some problems in the process, firstly some of the results were inaccurate due to under-representation of the data in this training. Secondly, due to the quality of the data, it may have made some serious errors in sensitive questions, such as background, light and other factors, as well as some stereotypes that made it biased and make it react in a biased way. I believe it would have been more meaningful if it had been given a more comprehensive dataset in subsequent attempts. So quality data is essential. Also be careful not to let it fall into unconscious bias, so building a more sensible code framework is necessary, but of course as of now it's beyond our capabilities.

Week 8

This week's topic is identity, algorithmic identity and data. I was surprised to find that Google was able to accurately predict my age, level of education and other more private data, as I was under the impression that I hadn't provided it with similar information, which made me realise that I was being subjected to a great deal of digital surveillance without even realising it. Even though I agreed to Google's user privacy agreement at the very beginning, I honestly never read the contents of it carefully because it was very long and the language was very obscure. In the data collection after the class, first of all I admit that it was very difficult for me because I don't have that many friends and even some of them never post content, so I lowered the requirements as long as we follow each other in Douyin as well as know each other to a certain extent. I also had some problems categorising the content they posted, for example my friend who is a photographer posted a gallery of geese on his page, should this be in the work category or the animal category? Also I found that the content that these friends of mine posted in their personal pages had a very distinctive feature, the content tended to focus on one or more categories. Finally I think that the categories provided in the table are not sufficient, for example my friends post Surreal antics, which I don't think would be comprehensively categorised as jokes, and photos of them travelling with local buildings and residents, which don't seem to make sense if they are just categorised as outdoors.

Week 9

This week, I attempted to analyze my friends using the data obtained last week. Given the large number of categories, I referred to SDT (Self-Determination Theory) for better visualization. According to this theory, I grouped the 15 original categories into three overarching needs: autonomy, relatedness, and competence. For each friend, I created a pie chart that illustrates the proportion of these categories, providing a clear and intuitive view of which aspects are more prominently represented in their online presence.In subsequent interviews, I learned that my friends generally believe the content they post can, to some extent, reflect their interests in specific areas, although the degree of representation varies. Therefore, I think it is essential to also consider their interactions with social media—such as likes, comments, and shares—as these behaviors might offer a more direct indication of their preferences.Compared to traditional media, personalized content recommendations clearly deliver content more aligned with each user’s preferences. Advertisements interspersed within these recommendations can also achieve better conversion rates. However, this approach also comes with potential downsides, such as algorithmic bias, the formation of echo chambers, and privacy concerns.

Week 10

This week's theme is Interactive Narratives, which reminds me of my previous experience creating H5 (HTML5) web pages. The two are quite similar. Interactive storytelling can make the narrative dynamic, enhancing user engagement and making it easier for users to immerse themselves in the story. It also allows the narrative to proceed according to the user's thoughts (or expectations). While the outcome might sometimes surprise the user, this unpredictability is part of the fun. This form of interactive storytelling is widely used in today’s gaming industry, such as the recently popular PC game Black Myth: Wukong. Players can interact with NPCs through dialogues to unlock different story endings or even access hidden levels. Finally, when designing interactive storytelling, it is important to consider cultural differences, gender representation, and social contexts to cater to a broader audience and provide more diverse options for users.

+++XUBIN WANG+++

Thank you for watching!!!