CV & Recommendations

The latest copy of my CV can be found here

A list of recommendations provided for me can be found here

Project List

Click on titles of projects for more details!

Taylor Swift Telegram Bot (2023)

Created a Telegram bot (@tswift_lyrics_songs_bot) that provides a random Taylor Swift song recommendation, including a Spotify link to the song(/random) and also a way to select what Taylor Swift eras fans should dress up as (/eras). For /eras, you select the album(s) that you’re willing to dress up as through a poll and the bot would select one for you, while providing a sample image of what the outfit’s style is like.

Tools utilized: Python (Telegram and Spotify APIs)

Netflix Top 10 (2022)

Created a website as an avid Netflix fan that’s often looking for the next thing to watch. This improves on the Netflix Top 10 website. Users can choose which country they want to search for and it’ll be produce a side by side list of the Top 10 Movies and Top 10 TV Shows in that country. Instead of just the film/show title that’s provided in the original Netflix Top 10 website, this website also produces the plot summary, IMDB rating and trailer. I hope to provide fellow avid Netflix fans with more information to help them choose what to watch next!

Tools utilized: Javascript, HTML, Python (Flask and BeautifulSoup)

Belay (Slack Clone) - Web Development (2021)

Created a website (Belay) that acts as a Slack clone with several of Slack’s key features. Website was built on React and all user information, channels, messages and threads were stored in a MySQL database which was accessed via Flask APIs. Users could sign up with their email and join channels to chat with others along with other features.

Tools utilized: React JS, JavaScript (JSX), Python (Flask) and MySQL

Gender Inequality - Data Vizualization (2021)

Produced 2 interactive data visualizations exploring gender inequality, in particular, the gender wage gap that persists in the US. The first allows users to compare the US’s performance to other OECD countries while the second visualization allows readers to explore gender wage gap across and within different industries in the US.

Tools utilized: Python (Altair), JavaScript (D3) and HTML

Big Data Architecture - FIFA 07 to 21 (2020)

Deployed a serving and speed layer (based on Lambda Architecture) via AWS load balanced servers that allows users to search for their favorite Football (Soccer in the US) team. Will return the team’s aggregated stats in different editions of the FIFA game. This allows fans to assess if their favorite or most hated teams have improved or deproved over the years and in which categories the team has changed.

Tools utilized: AWS S3, EMR, Spark, Hbase, Hive, Hadoop, Kafka Message Queue, HTML and CSS

NBA Players iOS App (2020)

Created an iOS app that allows users to search for NBA players and return key statistics of selected players. Users are also able to save players into their ‘Favorite Players’ list or move players into an ‘Injured Players’ list. The app is meant for fans of the NBA who are curious about certain NBA players or for fantasy basketball players who need to be kept up to date on the stats of different NBA players.

Tools utilized: Swift in xCode

Which area of Singapore has the best restaurants? (2020)

Dividing Singapore into 55 planning areas or 332 subzones, I created an interactive visualization using the Bokeh library in Python where users can explore which areas have the best and most popular restaurants in Singapore. Restaurant data came from a TripAdvisor crawler that I wrote as well as Yelp data kindly provided by Sean Koh.

Tools utilized: Python via Juypter Notebooks

Machine Learning - Predicting Occurence of Violent Crime in Chicago (2019)

Using socio-economic, weather and reported crimes data, we attempt to assess the possibility of using a ML classification model to accurately predict if a violent crime would occur in a specific neighborhood in Chicago. This seems particularly important given recently controversies around police brutality and calls to ‘defund’ the police which makes more efficient allocation of limited police resources more important. We will be identifying the ‘best’ combination of hyper-parameters in 4 different models: decision tree; random forest; logistic regression & ada-boost that produces the most accurate predictions.

Tools utilized: Python via Juypter Notebooks

Predicting Success of Android Mobile Apps (2018)

Attempting to establish what factors are important for app success on the Google Play Store so that we’re able to predict if an Android mobile app will be successful. Applied 3 different classification modeling approaches - Logistic Regression, Decision Trees/Pruned Decision Trees and Random Forest models. Found that the random forest model was most successful and that the variables - Price, Rating, DaysSinceLastUpdated and Size are most important in prediction success.

Tools utilized: R

Divorce Trends in Singapore (2018)

Attempt to examine Divorce and Marriage trends in Singapore with a particular focus to divorce trends which has been undeniably on the rise around the world. Singapore’s crude divorce rate remains low relative to other countries in the West such as the US and Russia but even among other Asian countries such as China. However, there are still some interesting trends in Singapore where the number of divorces increasing by 5x from 1980 to 2018 despite the number of annual marriages remaining stable.

Tools utilized: Tableau

Exploring Racial Inequality in Education for Singapore (2018)

Using 2005 and 2010 data, I explore how racial inequalities in educational outcomes is a serious problem. My goal is to demonstrate that Malays, relative to other ethnic groups in Singapore, are badly under-represented in higher (tertiary education) though improvements have been made. I also attempt to show that this is exacerbated by the fact that Malays who make it to University are relatively under-represented in high paying fields such as Engineering and Law.

Tools utilized: Tableau