Hey there 👋

I'm Khin Yadanar Lin

Data Science & Analytics

Results-driven data professional with 8 years of experience in quantitative research, analytics, and management consulting. I am passionate about delivering impactful business solutions through the power of big data technologies and data science techniques.

Let's connect! View projects on Github

Featured Projects

Topic Modelling & Sentiment Analysis - Data Science Subreddit

I crawled over 7,000 top posts and comments from the 'Data Science' subreddit using the Reddit API. I then applied text preprocessing and TF-IDF vectorization before training an LDA model to extract five key discussion themes. I also employed a lexicon-based sentiment analyser to guage the sentiment across different flairs.

Python PRAW NLTK Tf-idf Vectorizer LDA Vader

False Fire Alarms Detection - London Fire Brigade

In the UK, 42% of emergency incidents reported to rescue teams are false fire alarms, costing approximately £1 billion annually to address. In this project, I analysed spatial data using geopandas, and built a classifier to detect false alarms using techniques like Logistic Regression, KNN, Naive Bayes, and Random Forest.

Python geopandas sklearn Classification Supervised Machine Learning

Car Auction Price Predictions

In this freelance project, I developed a ML pipeline to predict used car prices for a Myanmar-based car rental platform. I identified key factors influencing resale values through feature importance analysis and employed techniques such as Ridge Regression and Ensemble Methods. The model is demonstrated via a Flask web app.

Python Statsmodels sklearn Flask Supervised Machine Learning

Inquiry Volume Forecasting - ARIMA Time Series Forecasting

This project uses real-world business data from an anonymised company to forecast daily chat inquiry volumes on Facebook Messenger through ARIMA modelling. Specifically, I explored dynamic regression with ARIMA error terms method, to allow the inclusion of additional predictors such as digital ads spend and campaign data

R ARIMA Time Series Forecasting Dynamic Regression

View more projects

Latest from Medium Blogs

Optimisation: Unpacking Queueing Theory in its Simplest Terms

Have you ever found yourself waiting in line at a supermarket, a restaurant, or a bank, wishing your turn would come just a bit faster?

Lessons From My ML Journey: Data Splitting and Data Leakage

Common mistakes to avoid when you transition from statistical modelling to Machine Learning