Enoch Loy

Intro

Welcome to my data science portfolio! I'm thrilled to share my journey of transitioning from a pharmacy background to a proficient data scientist with a passion for learning. My track record showcases my ability to craft precise models and extract actionable insights even within narrow timeframes.

My experience thriving under pressure in the fast-paced pharmacy environment and intensive bootcamp equips me to meet tight deadlines without compromising quality, a skill I've successfully applied to my data work.

I'm actively seeking opportunities to use my skills to solve data-driven problems. As you explore my portfolio, envision how my unique background and data expertise can create a positive impact.

If you are looking for a dedicated data professional with a relentless appetite for learning, let's connect and explore how I can add value to your team. I'm eager to turn data into results.

Machine Learning Projects

Skin Disease Classification using Computer Vision

In light of the prevalence of skin diseases and their self-treatable nature, accurate classification of common skin conditions can significantly contribute to addressing healthcare burden.

I developed a robust image classification model utilizing deep transfer learning to effectively classify five prevalent skin conditions in Singapore, namely acne, eczema, fungal skin infections, psoriasis, and warts. This model has showcased impressive accuracy, achieving above 90% on all key metrics.

I deployed the model on Streamlit, providing a user-friendly platform that offers a preliminary diagnosis and valuable information on the diseases and their management.

Key Skills: Python, Tensorflow, Keras, Convolutional Neural Networks, Transfer Learning, Deep Learning, Grad-CAM, Web Scraping, BeautifulSoup, Data Augmentation, Ensemble Modeling, Data Visualization, Docker

GitHub Repository

Live Demo on Streamlit

Tackling the West Nile Virus Outbreak with Data

Leveraging mosquito population data and weather conditions, I developed a binary classification model to predict the presence of West Nile virus in mosquitoes across Chicago.

My model achieved ROC AUC score of 88%, effectively predicting West Nile virus presence in traps. Prioritizing areas for pre-emptive spraying can help to optimize pesticide deployment and minimize costs while safeguarding public health.

Key Skills: Python, AdaBoost, CatBoost, GradientBoost, XGBoost, Logistic Regression, Random Forest, Regularized Greedy Forest, Data Visualization, Data Analysis, Machine Learning, Predictive Modeling, Binary Classification

GitHub Repository

Live Demo on Streamlit

Reddit Binary Classification for Diet Communities

I utilized web scraping and natural language processing (NLP) techniques to collect and analyze a dataset comprising of over 2000 user posts from Keto and Paleo diet communities on Reddit.

The binary classification model achieved 94% accuracy and 93% f1-score in predicting subreddit origins based on user posts. Through deep analysis of user-generated content such as bigram/trigram words, I also identified unique patterns which can guide targeted marketing and product development.

Key Skills: Python, Natural Language Processing, Web-Scraping, Logistic Regression, Multinomial Naive Bayes, Bernoulli Naive Bayes, Gaussian Naive Bayes, K-Nearest-Neighbors, Count Vectorization, TFIDF Vectorization, Binary Classification, Machine Learning, Data Analysis

GitHub Repository

Live Demo on Streamlit

Predicting HDB Resale Prices

HDB flats are highly sought after due to their affordability and amenities, making accurate price predictions vital for both buyers and sellers.

Using a comprehensive dataset from a Kaggle competition, I developed a linear regression model to predict the resale price of HDB flats. For simplicity, I reduced the number of features from 77 to 16, and achieved a model R² score of 0.9 and root-mean-squared error (RMSE) of $46,000.

Key Skills: Python, Data Cleaning, Exploratory Data Analysis (EDA), Feature Engineering, Linear Regression, Ridge Regression, Lasso Regression, ElasticNet Regression, Data Visualization, Data Analysis

GitHub Repository

Data Analysis Projects

COVID-19 Data Analysis and Visualization

I conducted an extensive COVID-19 analysis using SQL, uncovering vital insights into cases, deaths, and vaccinations. This project encompasses data preparation, exploratory analysis, and in-depth examinations at global, continental, and country levels. SQL techniques utilized include data loading, aggregation, subqueries, common table expressions (CTE), temporary tables, creating views and window functions.

Additionally, I created an interactive Tableau dashboard, which features:

• Geographical map showing population percentages for vaccinations, deaths, and infections.
• Area chart illustrating rolling totals for vaccinations, cases, and deaths over time.
• Treemap for visualizing the probability of death if infected by country and continent.
• Bar chart identifying the top 5 countries with the highest population percentage of deceased.
• Line chart tracking new cases over time.

Some key insights include:

1. A seasonal component exists for COVID, with a spike in cases towards the end of the year.
2. South America has the highest vaccination per capita (1.9), yet they have the highest death per capita (0.0031). Asia has the lowest cases and deaths per capita compared to other continents.
3. Yemen has the highest death probability if infected at ~18%, a stark difference from the second highest country Sudan at 7.9%.

Key Skills: SQL, MySQL, Exploratory Data Analysis, Data Cleaning, Extract-transform-load (ETL), Data Integration, Data Analysis, Data Visualization, Tableau, Interactive Dashboards

GitHub Repository

Tableau Public

Climate Data Analysis for Picnic Planning

I completed a comprehensive analysis of climate data specific to Singapore, aiming to optimize picnic planning and operations by leveraging weather insights. Utilizing Python and the Pandas library, I cleaned, validated and integrated data from 8 sources into an unified view before conducting exploratory data analysis and visualization to uncover key insights. Visualizations used in this project include heatmaps, histograms, boxplots, scatterplots and line charts.

Some key insights include:

1. February has ideal conditions for a picnic, such as best wind speeds and lowest rainfall.
2. Rainfall variation by weekdays exist: heaviest rains on Friday, highest probability of rain on Wednesdays.
3. Changi offers a higher probability of favorable weather conditions compared to Tuas and Ang Mo Kio.

Key Skills: Python, Extract-transform-load (ETL), Data Cleaning, Data Integration, Data Validation, Pandas, Numpy, Data Transformation, Exploratory Data Analysis, Data Visualization, Matplotlib, Seaborn, Statistics, Data Analysis

GitHub Repository

About Me

I graduated with a Bachelor's Degree in Pharmacy from National University of Singapore, and worked as a hospital pharmacist for 2.5 years thereafter. Venturing beyond my comfort zone, I used my spare time to pick up essential skills in finance, investing, programming, data analytics and data science. As a person who loves solving difficult problems, data challenges were akin to puzzles, and I was soon drawn to this exciting domain.

I eventually made a pivotal decision to enroll in General Assembly's intensive data science bootcamp, where I honed my expertise through a diverse array of projects. My data journey has only just started and I am excited to see how it unfolds.

Skills

Contact

Find me at:

91371291
loyenoch@gmail.com

• Linkedin
• GitHub

Elements

i = 0;

while (!deck.isInOrder()) {
    print 'Iteration ' + i;
    deck.shuffle();
    i++;
}

print 'It took ' + i + ' iterations to sort the deck.';