The Missing Blog Post
Funny enough, I've studied algorithms in depth. From how a simple mathematical algorithmic sequence works, to applications in nature and in computers, this part of my life has been belabored until I could recite a simple definition and example in my sleep (please don't make me I'll cry). In laymen's terms, an algorithm is a logic based set of instructions that allows for a program's expression of a "semantic property." This property, in terms of a Google search would be the search results, with the instructions being your search.
My research question was simple, under the influence of different users, how does the Twitter "for you" algorithm change? I controlled for geography and time accessed, but changed sex/gender of individuals asked. I found that the "for you" page was not as heavily influenced by individual influences, as both my list and a friend of mine with a completely different set of likes had an overlap of 5 topics that both of us were not interested in. These topics, however, were popular in our given area, with our given age group, and therefore were pushed toward our trending page.
This information does show that specific hashtags are important, but also keywords. Twitter doesn't just pick up on the use of hashtags (since many users don't use them). This can be useful if an individual wanted to post their portfolio to Twitter. Monitoring the Trending/For You sections could give keywords that need to be included in the description in order to effectively have the website ping more individuals.
It must've been interesting to dive into algorithms in such a simplistic way after researching them as in-depth as you have. As someone who *hasn't* researched them before, I find myself surprised that one's sex doesn't influence their results more. The age/location/keyword factors definitely make sense, but I feel like across the board men's results, especially within social media, differ dramatically from womens' (in cases where you choose M/F). I have two sisters and we have created accounts at the same time with the only difference being a year or two apart for DOB and sex, and the FYPs are very different. I wonder if this has changed in recent years!
ReplyDeleteThis is such an intriguing blog post! It's fascinating to read about your in-depth study of algorithms and their applications. Explaining algorithms in laymen's terms and using the analogy of a Google search is a great way to make it relatable. Your research question regarding the Twitter "for you" algorithm and how it changes under different user influences is very interesting. It's impressive that you controlled for geography and time accessed while considering the impact of gender. The discovery of common topics in the "for you" page, despite different sets of likes, highlights the influence of local trends and age groups. Your insight about the importance of specific hashtags and keywords on Twitter is valuable, especially for individuals looking to promote their portfolios. Monitoring the Trending/For You sections to identify relevant keywords is a clever strategy. Thank you for sharing your findings and knowledge!
ReplyDelete