### Status: 🟢 Online

**Made with Angular**

## [View the site](https://yukseltron.github.io/Josephus-Problem/)

```cv
gallery
{"layout":"hscroll","images":[{"src":"https://firebasestorage.googleapis.com/v0/b/maitake-project.appspot.com/o/pages%2F6KQG3gmhSra20VnXw3yALJ7M20J3%2F3I0GxLCNCCsugnmNQs4m%2F1971257a-068e-4dd3-8c25-bd4190540be4.gif?alt=media&token=e2d0ddd5-36bc-4e14-a1dc-d3aa83a8e354","id":"1971257a-068e-4dd3-8c25-bd4190540be4","width":1316,"height":1190,"filename":"jp.gif","type":"image/gif","caption":"","border":false},{"src":"https://firebasestorage.googleapis.com/v0/b/maitake-project.appspot.com/o/pages%2F6KQG3gmhSra20VnXw3yALJ7M20J3%2F3I0GxLCNCCsugnmNQs4m%2Fe83da13e-21e0-4593-8b62-42a1ec2932df.gif?alt=media&token=8332bda1-1d23-45cd-a783-b6666df034d3","id":"e83da13e-21e0-4593-8b62-42a1ec2932df","width":600,"height":411,"filename":"3.gif","type":"image/gif","caption":"","border":false}]}
```

This is a simple visualizer of the Josephus Problem. This was really just me trying out [Angular animations](https://angular.io/guide/animations). Part of the trick here was creating an evenly spread out circle of circles that could adjust to various sizes.

The Josephus Problem

**Role:** Junior Web Developer

**Timeline:** Summer 2017

**Location:** Hamilton, ON ⛰

```cv
gallery
{"layout":"hscroll","images":[{"src":"https://firebasestorage.googleapis.com/v0/b/maitake-project.appspot.com/o/pages%2F6KQG3gmhSra20VnXw3yALJ7M20J3%2FpKaV7TVr8bwXdSzqBqUJ%2F94e49610-26c0-4e4c-94fe-9a197003ca72.png?alt=media&token=03731c49-7fa8-41f8-83db-8b6e82644eb6","id":"94e49610-26c0-4e4c-94fe-9a197003ca72","width":748,"height":378,"filename":"Screen Shot 2022-01-14 at 12.05.38 PM.png","type":"image/png","caption":"","border":false}]}
```

## Background

This job I found interesting as it involved working more with newer web technologies and provided more experimentation work than my previous position. The SSC (Student Success Centre) focused on building projects meant to help McMaster students with their academic and professional success. This was a nice way to give back to the community and work on more proof-of-concept experimental work.

## What I did

My responsibilities mainly involved working on a proof-of-concept risk assessment app for student-led events. The idea was for automating risk assessment forms in order to reduce wait times that could last for as long as 2 weeks. By automating this, we could potentially reduce the wait times to at least the same day the form was submitted. Using Vue.js and Laravel, I coded a decision tree for risk assessment and implemented it on a web app. The risk assessment decision was based on factors such as event location, number of participants, event length, and expected activities.

## Outcomes

* Proof-of-concept app could potentially reduce wait times from 2 weeks to 1 day
* Successfully automated most tasks to reduce the amount of manual human work needed for risk assessments

## What I learned

* Vue.js and Laravel

* Proof-of-concept work

* Experimentation

* Task automation

* More web development!

## Reflections

While the risk assessment app was able to work as intended, it ultimately was not put into full-scale use. This is mainly due to the difficulty of automating risk assessment for events. Events often vary in scope and in potential risk. Some events for example might require students to sleep over for a 24h period (such as for a hackathon) and requires knowledge of building fire codes for approval. As a result, it was tough to automate risk assessment for events that featured complex activities. While most events were mainly talks or large gatherings, some events may include heavy equipment and security personnel.

At this point, a human would need to come in to better determine the risk for more complex events. However, a problem that emerged was for validating the event with what was written on the event risk form. It was essentially an honour system that trusted the form submitter was not lying or omitting important information to avoid revealing the event posed too high of a risk. This ultimately is what kept the app from moving on past a proof-of-concept stage. Careful auditing of event risk still needed human intervention. Regardless, working on this project taught me how to use more modern web development tools and in requirements gathering.

@ SSC

**Made with Python and Pillow (Python Imaging Library)**

## [View repo](https://github.com/yukseltron/scribbles-art-game)

```cv
gallery
{"layout":"hscroll","images":[{"src":"https://firebasestorage.googleapis.com/v0/b/maitake-project.appspot.com/o/pages%2F6KQG3gmhSra20VnXw3yALJ7M20J3%2F8EOowVYe1j2AMoZ2polV%2Fcee01bf2-895a-495a-a6c8-791d1307ded0.gif?alt=media&token=965a2253-f0b6-473a-b3a7-6d40fae4ad3d","id":"cee01bf2-895a-495a-a6c8-791d1307ded0","width":250,"height":250,"filename":"1.gif","type":"image/gif","caption":"","border":false},{"src":"https://firebasestorage.googleapis.com/v0/b/maitake-project.appspot.com/o/pages%2F6KQG3gmhSra20VnXw3yALJ7M20J3%2F8EOowVYe1j2AMoZ2polV%2Fb72c5f61-4125-4409-a357-7897d644c80a.gif?alt=media&token=69eb8c57-55dd-466b-92a8-3471ebfde75b","id":"b72c5f61-4125-4409-a357-7897d644c80a","width":1384,"height":794,"filename":"2.gif","type":"image/gif","caption":"","border":false}]}
```

This was a fun foray into image creation through code. Python's PIL is amazing at manipulating images, and the Pillow fork allows it to be even easier to implement.

As well, [I was inspired by this project](https://www.freecodecamp.org/news/how-to-create-generative-art-in-less-than-100-lines-of-code-d37f379859f/), in which pixel "video game enemies" are randomly generated. Of course, I wanted the images to look much different, so I took the approach of creating "scribbles" through the use of random shape outlines and vectors overlapping each other with random gradient colours. Doing this again with the same colour would create a second similar image to then create the GIFs you see above. \
\
Extending off of this, it was then added to a simple "shotgun" game to truly make it a video game boss, complete with a randomly generated name and quote.

Generative Scribbles Art

**Role:** Front-end Software Engineer Co-op

**Timeline:** Summer 2021

**Location:** Toronto 🦖

```cv
gallery
{"layout":"hscroll","images":[{"src":"https://firebasestorage.googleapis.com/v0/b/maitake-project.appspot.com/o/pages%2F6KQG3gmhSra20VnXw3yALJ7M20J3%2FJAlwlt27VWAUdupsOzca%2F1b87fb71-a919-48b5-8f7c-3a2154db0f37.png?alt=media&token=8c8fabb3-5254-4397-8c60-3457d38707f8","id":"1b87fb71-a919-48b5-8f7c-3a2154db0f37","width":764,"height":231,"filename":"league.png","type":"image/png","caption":"","border":false}]}
```

## Background

At some point, I started being more conscious of where I wanted to work next. So far I had mostly worked in fintech but I felt I should try to focus on a different industry. Healthcare was interesting as I liked the idea of working somewhere that took this as their main priority. League's focus in this area made it very appealing to me and I am happy to have had the opportunity to be there.

## What I did

At League, I was part of the engagement/rewards team (it was renamed from engagement to rewards mid-way). My main responsibilities involved working on refactoring and modularizing various web components on the main League web app. Specifically, I worked mostly on components for the home feed and user profile. Refactoring work involved updating the code to utilize League's atomic design system called Genesis and implementing dynamic code where possible. I took particular interest in Genesis as this was another design system I wanted to work on. League was still developing Genesis and an area I looked into was helping make Genesis semantic tokenization uniform and consistent across platforms and applications. This came with big obstacles and so I set up meetings with lead designers and engineers (on web and mobile) for discussions on understanding Genesis, its current implementation, and future goals. Bringing over my previous experience working with CaseWare's design system, I devised a project plan for milestones in implementing and maintaining consistency for Genesis tokens. As well, I worked with designers on potentially implementing a Figma plug-in for design linting.

## Outcomes

* Refactored and modularized components of the League web app home feed and user profile to be dynamic, scalable and accessible, greatly helping League meet accessibility compliance standards.
* P.o.C. work on implementing and ensuring design system tokenization consistency across platforms, allowing for more unified sources-of-truth which could help save significant engineering and design implementation/testing time. Independently set up meetings with lead designers and developers to work on requirements gathering and drive the project.

## What I learned/Improved on

* React

* Atomic Design Systems

* Contributing to large codebases

* Project planning and initiative work

## Reflections

```cv
gallery
{"layout":"hscroll","images":[{"src":"https://firebasestorage.googleapis.com/v0/b/maitake-project.appspot.com/o/pages%2F6KQG3gmhSra20VnXw3yALJ7M20J3%2FJAlwlt27VWAUdupsOzca%2Fab7bbcb5-8d1c-4767-a748-f39f9f37ca99.png?alt=media&token=3ecc14cc-65a9-444d-a8ce-070e42861b60","id":"ab7bbcb5-8d1c-4767-a748-f39f9f37ca99","width":1452,"height":776,"filename":"Screen Shot 2022-01-14 at 3.32.17 PM.png","type":"image/png","caption":"","border":false}]}
```

League was a very fulfilling place to work at. I think they are a company that really focuses on their culture and work-life balance. As well, working with the rewards team was a wonderful time! These were very talented people that I got the pleasure of learning a lot from. As well, I appreciate the maturity of how they treated me. Despite being a co-op, I feel I was really allowed to explore various areas and things to work on. This is what helped me get involved with Genesis and I'm grateful for how openly I was able to discuss it with other engineers and designers.\
\
Looking back, I think the biggest difficulty faced was dealing with the pandemic. Things were all done remotely and this was my first team working in a company this away. While it was tougher to communicate, I appreciate having a team that was open to having a lot of 1:1s. I think this was important in not just getting to know everyone, but also in being able to just talk with each other. Looking forward I can definitely say working at League set a higher standard for the type of work and environment I want to be in.

@ League

**Role:** UX Designer/Developer Co-op

**Timeline:** May 2018-2019

**Location:** Toronto 🦖

```cv
gallery
{"layout":"hscroll","images":[{"src":"https://firebasestorage.googleapis.com/v0/b/maitake-project.appspot.com/o/pages%2F6KQG3gmhSra20VnXw3yALJ7M20J3%2FHMNG26TbkPs10zL9ZGXv%2F1d3b3676-b394-46b7-a210-7286690a3148.jpg?alt=media&token=e54ab5f3-8a0b-4c95-a5db-15ae81e803b8","id":"1d3b3676-b394-46b7-a210-7286690a3148","width":720,"height":72,"filename":"CaseWare-Logo-RGB-Primary_Dark-TM 1.jpg","type":"image/jpeg","caption":"","border":false}]}
```

## Background

After my previous position that involved working closely with back-end technologies, I wanted my next position to be closer aligned to my interests in design. This was a bit of a pivot in what I wanted to do. I still wanted to get better at coding and software development, but I also wanted to pursue my increasing interest in design. Luckily I was able to find an opening at CaseWare where I could be both a designer and developer for their UI and design systems. I felt the design team at CaseWare were really smart people who could definitely help me grow in my career.

## What I did

There was definitely more I was able to contribute to thanks to being at CaseWare for a year. As a UX designer/developer for a small team, I was happy I could be put in various different projects that required different skills. Sometimes I would need to code something, other times I would need to design. The main responsibility I had was on helping build [NOVA](https://nova.caseware.com/), CaseWare's design system. I contributed to and helped maintain the style guides, and helped design/build the content featured. As well, I provided the illustrations seen on the site as well most of the icons in the icon library. Another word for what I'd call my role was "UX Engineer".

## Outcomes

* Successfully launched NOVA from beta and helped build/maintain it to version 2.0
* NOVA has been adopted by all of CaseWare's product teams
* NOVA has replaced a significant amount of UI, including 2 major cloud platforms and 3 other side products
* NOVA is extensively used for experimental and custom apps within the platforms.
* NOVA has helped meet A11y compliance nearly across the board

## What I learned/Improved on

* Web UI component development

* Design systems

* Microinteractions

* Web Accessibility standards

* Adobe XD, Ai

* JIRA

## Reflections

```cv
gallery
{"layout":"hscroll","images":[{"src":"https://firebasestorage.googleapis.com/v0/b/maitake-project.appspot.com/o/pages%2F6KQG3gmhSra20VnXw3yALJ7M20J3%2FAOhoc1FVzXQ2OuYxOmZs%2Ff7c1363a-0b3c-4738-9c6d-d2643329883f.jpg?alt=media&token=b009153f-a5ea-410c-a74e-cf23a193e71c","id":"f7c1363a-0b3c-4738-9c6d-d2643329883f","width":4032,"height":3024,"filename":"IMG_1339.jpeg","type":"image/jpeg","caption":"","border":false}]}
```

I am incredibly appreciative of being allowed to work on the NOVA design system. This is the first design system I helped build and I learned immensely from it thanks to the smart people I worked with. We were a small but inter-disciplinary team and would collaborate on multiple projects. So much of what I know today I can still credit back to my time working on this team. Those amazing people set the benchmark for me on all future group work.

There was so much I managed to do and learn, so I'm very grateful to the design team. They took me on and let me work on projects that really interested me. I feel I improved both as a software developer and a designer but also made some lasting friendships. While the design team has since moved on to bigger and better things, it was an incredibly formative time for me and I will cherish the memories.

@ CaseWare

### Status: 🔴 Offline

**Made with Python Django,  IBM Watson Tone Analyzer, Aztro astrology API, originally hosted on Google Cloud.**

## [View the repo](https://github.com/yukseltron/emoji-horoscope)

```cv
gallery
{"layout":"hscroll","images":[{"src":"https://firebasestorage.googleapis.com/v0/b/maitake-project.appspot.com/o/pages%2F6KQG3gmhSra20VnXw3yALJ7M20J3%2FJuL71zWht8StqLsaz2S7%2Fc4275213-9f02-4178-a092-e7d109cfc606.png?alt=media&token=e373f9da-be35-44b7-b296-c31623110db3","id":"c4275213-9f02-4178-a092-e7d109cfc606","width":1172,"height":516,"filename":"Screen Shot 2022-01-22 at 10.03.07 PM.png","type":"image/png","caption":"","border":false},{"src":"https://firebasestorage.googleapis.com/v0/b/maitake-project.appspot.com/o/pages%2F6KQG3gmhSra20VnXw3yALJ7M20J3%2FJuL71zWht8StqLsaz2S7%2F8b47d0ab-2170-4ced-bfbb-be6c4084b4e3.png?alt=media&token=a576f57c-16cd-45b4-9e2a-11230668ae80","id":"8b47d0ab-2170-4ced-bfbb-be6c4084b4e3","width":1022,"height":1034,"filename":"Screen Shot 2022-01-22 at 10.03.27 PM.png","type":"image/png","caption":"","border":false}]}
```

The focus for this project was on trying out Django (a Python web framework) and two API's. The web app is pretty simple, users input their astrological sign and then this makes a call to the Aztro API for the text horoscope. This text is then fed to the IBM Watson Tone Analyzer, which I then map the results of to specific emojis.

```
joy = ['😁','😀','😂','😆','🤑','🤣','😊','😋','😍','😇'] anger = ['😡','🤬','😠','🤯','😈','😤']
fear = ['😱','😰','😨','😬','😳','😥']
sadness = ['☹️','😖','😞','😢','😭','😕','🙃','😔']
analytical = ['🤔','🤨','🧐']
confident = ['😛','😜','😏']
tentative = ['😒','😑','😓','🤭']
```

In reflection, I could definitely work on improving the mapping from tone to emoji. I think maybe it could be more nuanced. But this also depends on the list of tones IBM Watson provides. Of course, I could also explore other sentiment analysis libraries.

Emoji Horoscope

**Role:** Software Design Engineer Intern

**Timeline:** Summer 2016

**Location:** Seattle, WA 🏔

```cv
gallery
{"layout":"hscroll","images":[{"src":"https://firebasestorage.googleapis.com/v0/b/maitake-project.appspot.com/o/pages%2F6KQG3gmhSra20VnXw3yALJ7M20J3%2FoxPC0UktMAOpZToqi1Cl%2F2c2a0b88-52ad-44bc-9e2f-fcf01390f2f1.png?alt=media&token=b7827180-b77e-4c88-9d6f-b8706a90f387","id":"2c2a0b88-52ad-44bc-9e2f-fcf01390f2f1","width":2521,"height":620,"filename":"Concur - Partner Logo.png","type":"image/png","caption":"","border":false}]}
```

## Background

Concur was my first professional internship and I guess job in software. I had met one of their recruiters at a hackathon (MHacks at the University of Michigan) and he encouraged me to apply. I did and soon later did 3 separate interviews for 3 separate roles. These were among my first ever interviews for a software job and Concur was the largest company to have given me an interview so far. Fortunately, I was able to receive an offer for the 3rd interview position as a software design engineer intern. I was also lucky in that I only went to MHacks that weekend because a friend of mine ignored me when I said I didn't want to go and showed up outside my place anyway with a ride (thanks, friend).

## What I did

I was placed on a team within Research and Development. The team was nicknamed team "Ekipazh" (meaning Flight Crew 🛫 in Russian). For the time I was at Concur, the project we handled was on building a multi-threading solution for quickly searching through different relational databases. My main responsibilities included building the web app used for dynamically showcasing the results of the multithreaded search. I also helped out in planning and designing the multithreaded solution itself in Java. As a small project, I also did p.o.c. group work for [a bonusly ](https://github.com/yukseltron/favourly)slack bot.

## Outcomes

* Successfully reduced wait times for database searches from \~6-8 hours down to 10-30 minutes
* Dynamic web app successfully shows in real-time multi-threaded database results, including successes and failures

## What I learned

* JQuery, Java, Spring, Git (in a team setting)

* Introduced to concurrent systems

* Introduced to relational databases

* MVC pattern

* Dynamic web development

## Reflections

```cv
gallery
{"layout":"hscroll","images":[{"src":"https://firebasestorage.googleapis.com/v0/b/maitake-project.appspot.com/o/pages%2F6KQG3gmhSra20VnXw3yALJ7M20J3%2FoxPC0UktMAOpZToqi1Cl%2F3c74ebe5-23d5-47f0-92cf-e779e0129984.png?alt=media&token=feeecf57-cc69-4a59-b602-680d1028bfa4","id":"3c74ebe5-23d5-47f0-92cf-e779e0129984","width":1652,"height":1086,"filename":"Screen Shot 2022-01-14 at 2.11.38 PM.png","type":"image/png","caption":"","border":false}]}
```

I am grateful for the opportunity that Concur presented me. It was fun being able to travel to another city in another country and work in tech. I made friends I try to keep in touch with and appreciate the time I had. If I were to go back in time, I think I'd tell myself to be more confident and take greater advantage of my time there. Seattle was a great place to be, but I was still unsure of what I wanted to do as a career. Did I want to become a web developer? I was learning a lot of design in my spare time as well. Seattle was a place loaded with talented artists and designers. Through the people I met and the friends I made, I think I eventually realized I needed to incorporate more design in my career and life.

That being said, while I still learned a lot working at Concur, I also faced some imposter syndrome and was not sure how to handle that. Part of this resulted from working with advanced concepts such as concurrency and relational databases. I was a first-year CS student at the time and a lot of this I had not studied yet. While this is no fault to the team or Concur, it was a bit intimidating as a result and sometimes doubts sprang up over whether or not I was suited for this role. I was fortunate to have had a kind and understanding team that helped get me started and feel welcome.  Looking back, I wouldn't say I have any real regrets. In the end, it was a big milestone for my career and I did my best at the time to make the most of it.

@ Concur

### Status: 🔴 Offline

**Made with Discord.js (a Node.js module for the Discord API) and TheMovieDB API.**

## [View the repo ](https://github.com/yukseltron/buscemi-bot/tree/master)

```cv
gallery
{"layout":"hscroll","images":[{"src":"https://firebasestorage.googleapis.com/v0/b/maitake-project.appspot.com/o/pages%2F6KQG3gmhSra20VnXw3yALJ7M20J3%2FYRiaJI2aw8MJvyNBYXev%2F4784a401-b757-4480-830f-3163cca991c2.png?alt=media&token=928a55cf-fc79-4f0e-8c04-eff7697893f0","id":"4784a401-b757-4480-830f-3163cca991c2","width":1368,"height":498,"filename":"Screen Shot 2022-01-22 at 1.12.58 AM.png","type":"image/png","caption":"The unfortunate takedown request :(","border":false},{"src":"https://firebasestorage.googleapis.com/v0/b/maitake-project.appspot.com/o/pages%2F6KQG3gmhSra20VnXw3yALJ7M20J3%2FYRiaJI2aw8MJvyNBYXev%2Feec16230-8f03-40b3-bd36-ade8d2550e70.png?alt=media&token=25ae7e9f-e356-49b3-b5c3-476351c5501b","id":"eec16230-8f03-40b3-bd36-ade8d2550e70","width":1120,"height":1092,"filename":"Screen Shot 2022-01-22 at 1.09.30 AM.png","type":"image/png","caption":"","border":false},{"src":"https://firebasestorage.googleapis.com/v0/b/maitake-project.appspot.com/o/pages%2F6KQG3gmhSra20VnXw3yALJ7M20J3%2FYRiaJI2aw8MJvyNBYXev%2Fb4316759-1de2-4887-b387-ec3fbe166e0f.png?alt=media&token=58fe80fb-7ea5-473e-bd32-cb6aaf8ce08d","id":"b4316759-1de2-4887-b387-ec3fbe166e0f","width":1130,"height":584,"filename":"Screen Shot 2022-01-22 at 1.11.26 AM.png","type":"image/png","caption":"","border":false}]}
```

This project was me trying to make a discord bot. Using TheMovieDB API, I created a bot that emulates "Steve Buscemi". This bot would be able to fetch quotes, clips, films, and images of Steve Buscemi. Why Steve Buscemi? I'm not entirely sure why. Looking back the project seems ridiculous. In actuality, the bot can be adjusted to act as any actor listed on API, I just for some reason chose to make it Steve Buscemi as an example. Naively feeling proud of the project, I even submitted it to [Discord Bot List](https://discordbotlist.com/). This way, the Buscemi bot would be hosted on an external server. \
\
However, before I could get the bot listed, I received a takedown notice from TheMovieDB. It had been sent from a certain someone I can only assume is associated with Steve Buscemi. Somehow, they had discovered I was using his likeness. I was surprised at the time, though I guess I simply did not understand the legal aspects. I had thought because I was not trying to monetize anything, it'd be ok. Oh well. I hope Steve Buscemi does not hold negative feelings against me for this.

Buscemi Bot

### Status: 🟢 Online

**Made with Knockout and the Giphy API**

## [View the site](https://yukseltron.github.io/giphy-tv/)

```cv
gallery
{"layout":"hscroll","images":[{"src":"https://firebasestorage.googleapis.com/v0/b/maitake-project.appspot.com/o/pages%2F6KQG3gmhSra20VnXw3yALJ7M20J3%2Fduqstmp2sy2fmQSVqY3o%2F4c57d322-dc53-414d-a8e9-6493c51dafed.png?alt=media&token=5ac59b2b-e00a-4227-b811-99e23d468b4e","id":"4c57d322-dc53-414d-a8e9-6493c51dafed","width":1972,"height":1304,"filename":"Screen Shot 2022-01-22 at 11.04.18 AM.png","type":"image/png","caption":"","border":false},{"src":"https://firebasestorage.googleapis.com/v0/b/maitake-project.appspot.com/o/pages%2F6KQG3gmhSra20VnXw3yALJ7M20J3%2Fduqstmp2sy2fmQSVqY3o%2F5656b072-f66e-4bfd-bc45-f4b670fc8a65.png?alt=media&token=3994ec9c-239c-44bf-bfb6-4b35c33c1fea","id":"5656b072-f66e-4bfd-bc45-f4b670fc8a65","width":1972,"height":1304,"filename":"Screen Shot 2022-01-22 at 11.04.40 AM.png","type":"image/png","caption":"","border":false},{"src":"https://firebasestorage.googleapis.com/v0/b/maitake-project.appspot.com/o/pages%2F6KQG3gmhSra20VnXw3yALJ7M20J3%2Fduqstmp2sy2fmQSVqY3o%2Fad5c6b7c-56f7-4f61-8046-89bbf90c985b.png?alt=media&token=4272b285-4fd7-4f45-8f3f-258d493d283f","id":"ad5c6b7c-56f7-4f61-8046-89bbf90c985b","width":1972,"height":1304,"filename":"Screen Shot 2022-01-22 at 11.05.15 AM.png","type":"image/png","caption":"","border":false},{"src":"https://firebasestorage.googleapis.com/v0/b/maitake-project.appspot.com/o/pages%2F6KQG3gmhSra20VnXw3yALJ7M20J3%2Fduqstmp2sy2fmQSVqY3o%2F45740ddb-e0cc-4a40-acf8-dd8620f90e55.png?alt=media&token=e758bda2-ee77-4e49-9d7e-3d480600e1d1","id":"45740ddb-e0cc-4a40-acf8-dd8620f90e55","width":1972,"height":1304,"filename":"Screen Shot 2022-01-22 at 11.04.14 AM.png","type":"image/png","caption":"","border":false}]}
```

An early project of mine back when I was still new to web dev. At that time, all these new JS frameworks and libraries were all the craze. Ember, Meteor, Angular, Knockout, Backbone, Vue, React, it was fun going through them and trying them out. Though things have changed and now I'd probably just stick with using React. This is the only time I've used Knockout. I had heard about it from a hackathon I had been to, and a friend recommended I try it out.  At the same time, I had been interested in exploring the Giphy API, so things sorta fell into place as I just tried making something simple that combines both of them.

GiphyTV

**Made with Python and CatBoost**

## [View notebook](https://github.com/yukseltron/ML-lyric-classifier/blob/master/2179-proj.ipynb)

```cv
gallery
{"layout":"hscroll","images":[{"src":"https://firebasestorage.googleapis.com/v0/b/maitake-project.appspot.com/o/pages%2F6KQG3gmhSra20VnXw3yALJ7M20J3%2Fg5Dm12Yk91heDfXkyTtp%2Faad3f923-fa3e-4d2a-9723-2e226fd29270.png?alt=media&token=477f79d7-3222-4b11-9a23-fcdb14533530","id":"aad3f923-fa3e-4d2a-9723-2e226fd29270","width":1312,"height":1122,"filename":"Screen Shot 2022-01-23 at 8.24.11 PM.png","type":"image/png","caption":"","border":false}]}
```

This project was for a machine learning applications course. It was completed within a week and was the first ML model I've worked on alone. The project goal was the same for everyone: given a dataset of real music lyrics, classify them into one of three genres (Rock, Hip Hop, and Pop).

### The Story

I needed to build a classifier, but I lacked domain knowledge in lyric composition. While there may be beautiful patterns that help sort the lyrics into a genre, I do not know any of that. So, it was hard to do any useful feature extraction from the lyrics. Things like lyrics length and number of verses vary a lot so simply finding quantities would not be enough. With limited time to work on this, it'd be difficult to gain a deeper knowledge of lyrical composition.\
\
The challenge here was in processing the lyrics. It was a (long) text feature, and so as a categorical data type is a bit tricky to convert into a numerical feature for use in other regression methods. That's when I began to wonder if I really needed to work that hard in converting the categorical data to numerical. Surely, there must be something already out there to handle text features easily.

While exploring some ML libraries, I came across CatBoost. CatBoost employs gradient descent and was especially interesting for its focus on working with categorical data. A little digging deeper into the CatBoost docs (which was also a challenge), I found that CatBoost actually has support for text features. \
\
Implementing CatBoost on the lyrics was then super easy:

```
#Build the model
model = CatBoostClassifier(iterations=100,
                           learning_rate=3,
                           depth=6,
                           loss_function='MultiClass')
```

### The model's accuracy: \~68-70%

This accuracy turned out to be really good, comparatively speaking. But, I didn't realize how good when I submitted it. In fact, I had fumbled my first submission, using testing the model on the holdout set rather than the testing set to my embarrassment. So when I submitted this for a second time at 3 AM on the day before the deadline, I was really just hoping I at least had done ok.

### Draft 1 accuracy:

> **From:** ML Professor
>
> The ABSOLUTE distance between your predicted accuracy and the real accuracy is: 0.125\
> \
> This is quite a lot.

When I submitted my first draft for testing, it turns out I had made quite some mistakes. CatBoost has extensive docs but lacks in greater details of its capabilities. This can sometimes make it hard to understand what CatBoost can do and how. Regardless, after the poor results of the first submission, I had to decide on what my next approach would be.\
\
CatBoost was not explicitly taught in the course. The professor encouraged us to use libraries outside those he had demonstrated in class, so I was immediately eager to find something new. I don't think I could carefully explain to you how gradient boosted descent works, but nor did I need to for the context of the project. But still, here I was staring at my broken CatBoost model, wondering if I should throw it all out the window and use something more tried and true like K-Fold and Naive Bayes. Instead, I chose to keep going with CatBoost. I believed if I could research it more, I'd understand how to properly implement the text feature.

### Draft 2 accuracy:

> **From:** ML Professor
>
> The ABSOLUTE distance between your predicted accuracy and the real accuracy is: 0.001
>
> You have a very good accuracy, congratulations!

For the second draft, I looked further into CatBoost and outside the docs.  After watching [this very helpful video](https://www.youtube.com/watch?v=usdEWSDisS0&t=1s), I was able to find and fix the mistakes I had made prior. Checking the accuracy of the model, this time it looked much more promising! 

After submitting it for the final time, I later found out during a class that the submission with the highest accuracy was one that used CatBoost. There was also only one person apparently who used CatBoost, so I guess we can infer whose submission that was ;)

Lyrics Classifier

### Status: 🟢 Online

**Made with Vue.js and hosted on Vercel**

## [View the site](https://monodrift.vercel.app/)

```cv
gallery
{"layout":"hscroll","images":[{"src":"https://firebasestorage.googleapis.com/v0/b/maitake-project.appspot.com/o/pages%2F6KQG3gmhSra20VnXw3yALJ7M20J3%2Fgyxy8sMVmfwSUJKJADOy%2Fd66e4267-3884-456a-956b-4b5577711792.png?alt=media&token=08e3006a-013f-4fc7-b0d2-f5da01f0dab9","id":"d66e4267-3884-456a-956b-4b5577711792","width":2880,"height":1330,"filename":"Screen Shot 2022-01-21 at 11.59.44 PM.png","type":"image/png","caption":"","border":false},{"src":"https://firebasestorage.googleapis.com/v0/b/maitake-project.appspot.com/o/pages%2F6KQG3gmhSra20VnXw3yALJ7M20J3%2Fgyxy8sMVmfwSUJKJADOy%2Fbaccf2ca-7615-4709-b359-fcadb0da082d.png?alt=media&token=561b6c97-2af0-4821-955d-e8ce39df00ba","id":"baccf2ca-7615-4709-b359-fcadb0da082d","width":2874,"height":1328,"filename":"Screen Shot 2022-01-22 at 12.04.21 AM.png","type":"image/png","caption":"","border":false}]}
```

I'm a big sci-fi fan. And sometimes, that ends up with me trying to do some writing. Of course, I am not a talented writer and have little experience writing in long-form. So, I just wrote these super-short stories with a sci-fi flair to them. Apparently, I later learned from a friend this form of writing could be considered as [Flash Fiction](https://en.wikipedia.org/wiki/Flash_fiction), where the word count lies at around \~1000.

Anyhow, this project also involves a development side. I wanted to work on a small Vue.js side-project while also putting in some writing. The website was easy enough to make, the content? Not so much. Since I've only written a few of the stories (which looking back, I do think I could do better now), there are some that are unfinished.

I do have notes that I've written somewhere in some journal that details what I want the other stories to be about. So one day, I think I will come back to this and rewrite it with new lenses. Literally too. My glasses are new.\
\
With a whirr, the machine spits back out a small grey card.\
The card is chipped and bent at the corners.\
Parts of it still retain its gloss, while most of it is skinned, exposing thin strands of plastic bleached by the sun pealing off.

Mono-Drift

### Status: 🔴 Offline

**Made with Swift**

## [View the repo ](https://github.com/yukseltron/buscemi-bot/tree/master)

```cv
gallery
{"layout":"hscroll","images":[{"src":"https://firebasestorage.googleapis.com/v0/b/maitake-project.appspot.com/o/pages%2F6KQG3gmhSra20VnXw3yALJ7M20J3%2FvbS24hYQEyztEcR626ti%2Fe3057f79-a321-4850-b242-5308d30e886b.png?alt=media&token=d83a2790-6b42-47e7-b6cf-d939c4be349c","id":"e3057f79-a321-4850-b242-5308d30e886b","width":1168,"height":2096,"filename":"1.png","type":"image/png","caption":"","border":false},{"src":"https://firebasestorage.googleapis.com/v0/b/maitake-project.appspot.com/o/pages%2F6KQG3gmhSra20VnXw3yALJ7M20J3%2FvbS24hYQEyztEcR626ti%2Ffa203471-0858-40de-a0a8-33cb581ee249.png?alt=media&token=4b275a08-3830-48a4-885c-acf1294e30c8","id":"fa203471-0858-40de-a0a8-33cb581ee249","width":1168,"height":2096,"filename":"2.png","type":"image/png","caption":"","border":false}]}
```

This project was part of a university class working with iOS development. The goal was to build an app for helping you journal and track your house plants. While this is a common app, the difference here is that this app focuses not just on time since watered, but more importantly on the moisture of the dirt. \
\
After reading a very interesting book called *[The New Plant Parent](https://www.houseplantjournal.com/product-custom/the-new-plant-parent/),* it was interesting to learn the importance of maintaining different moisture levels for different types of plants. This, combined with the light level the plant receives, is far more important than simply rewatering it on a consistent basis. \
\
As such, this app would have a registry of common houseplants and provide the appropriate level of moisture and light the plant prefers. As well, a calendar was also implemented for tracking water days. 

\
While this app is not on the iOS store, it was a fun introduction to Swift but a brutal introduction to SwiftUI. SwiftUI had just come out around the time I started this project and I went ahead with trying to implement it. It was a bit tricky, to say the least. A newer tool lacking in appropriate documentation at the time, SwiftUI provided a challenge to implement. But... I can definitely see the potential in SwiftUI. I think in theory it does make the UI much easier to implement than before, so I do look forward to seeing it develop and mature.

```cv
gallery
{"layout":"hscroll","images":[{"src":"https://firebasestorage.googleapis.com/v0/b/maitake-project.appspot.com/o/pages%2F6KQG3gmhSra20VnXw3yALJ7M20J3%2FvbS24hYQEyztEcR626ti%2Fed67bd65-03c3-49ad-8738-3a121922ada8.jpg?alt=media&token=ac3bd93f-61a4-4a3d-ac4d-f4dc481cd747","id":"ed67bd65-03c3-49ad-8738-3a121922ada8","width":1149,"height":1532,"filename":"65679253891__CD89D953-C3FB-4D0E-9B87-EFD7461BA4E3.jpeg","type":"image/jpeg","caption":"The secrets are in the soil","border":false}]}
```