Lots of things happened in the last six months. After another academically successful term, I rented a VPS. I dug into Linux, PHP, MariaDB, CSS, JavaScript, jQuery, even Socket.IO. I learned many different things including image cropping, real-time private chat systems, collecting/processing geographical data, AJAX, and personality based user filtering in order to create a niche social network. It was a great full-stack web development project to spend my summer vacation. All of the hardest sub problems were solved and the website was nearly complete when the school started. Then, I could not find time to work with the rest. I'm planning to complete and release it during the semester break, in early 2018. I think it can fill a gap and come in handy, so I'm excited.

Update: I became a research assistant and I could only finish it in mid-2019. I am struggling to find new users, mainly because people do not trust since the website is a rogue project for now. I cannot focus on the website for now.

In other news, I gradually trimmed and finally shaved my beard. Now, I only have a soul patch along with a moustache and (consequently?) a girlfriend. I'm mentioning this because I was using Tinder and I decided to manually collect my data as I realized a pattern was going on. For example, I had a much higher chance to get matched with girls who had an extreme hair color such as purple, and I wanted to analyze the data. My suspicion was correct. At that time, I was also trying to learn machine learning, so I decided to apply my limited knowledge to the data I have in order to calculate the probability of a girl with specific profile attributes to like me given that I liked her. I initially used MATLAB, but I wanted to build an interactive web application. So, I found a JavaScript library to produce random forests and I used my dataset that consists of 550 girls I liked. Then, I also learned how to use Chart.js and CanvasXpress to create 2D/3D graphs. The thumbnail of this post actually belongs to the 3D graph I created with that application. I don't want to disclose much, because my knowledge was really limited. In the future, I might return back to this domain, do it better and write a blog post about it. For now, my Tinder adventure is happily over. I was sick of documenting the profiles, manually entering the data and having superficial conversations that did not get me much further. In the future, I would love to work at Tinder or OkCupid, though. Online dating is a fascinating subject.

Anyway, this term I'm taking a machine learning course, which is both exciting and terrifying. When I look back, I see a great progress. However, my long-term goals are ambitious and I'm still very far from that point. Implementing machine learning algorithms is valuable, but I need to do much more than that. I need to really understand and internalize it in order to be able to move it forward and make a difference. I'm worried. Will I ever get there? We will see. My term project for the course is about predicting a given tweet's interaction amount on Twitter. I've been also learning R to analyze social networks, because of another course I'm currently taking.

This term is busy as well. The courses are mostly easy to follow, but their projects and responsibilities need serious commitment, and this discrepancy stresses me out. By the way, my personality test application has been downloaded 2963 times and I've published a Chrome extension.