My Interest

• Statistics
• Machine learning/Deep learning
• Software/Web development

Always learning new technologies. GitHub is a space where I share my learning progress and learn from others. Also looking for collaboration opportunities.

---

- Languages

- Data Science, Machine Learning & Deep Learning

- Web Development

- Version Control

- Databases management

- OS

---

We look at the basics of implementing a neural network in regression and classification tasks. We cover basic concepts such as activations, loss and optimizers. We also go through common techniques such as learning rate scheduler, hyperparameter tuning and dropouts. We also include a section on image processing with Convolutional Neural Network (CNN). We implement a shared music controller where a host can create a room and connect his Spotify account to the room through the Spotify API. Other people can join the room through a room code and (depending on settings) can control (pause/play) songs. People can also vote to skip a song. Once enough vote is gathered, the next song is played. The app framework is built on django, with the help of rest-framework to handle database operations and connecting to Spotify API. Finally, React JS is used to implement the frontend. As a data analyst of a popular music store, we explore their database and provide business-relevant insights. Concepts such as union, join, aggregate and wildcards are reviewed. Detailed codes and results are provided for easy follow along. We also tackle more complicated tasks such as finding summary statistics from multiple tables, and creating temporary tables. We classify three species of puffins based on their attributes such as beak length and body mass. We will implement a standard machine learning workflow that covers

1. Data cleaning, wrangling, exploration and visualization
2. Model fitting and hyperparameter tuning
3. Model comparison and evaluations
4. Picking the best model

Popular classification algorithms such as K-means clustering, logistic regression and naive bayes classifier are reviewed. Performance metrics like accuracy, precision and recall are also discussed.

Here we explore ecology modelling concepts with R software. The Lotka-Volterra (LV) model describes the predator-prey dynamics in a natural habitat. In this project we aim to look at different factors and how they influence the populations. We also demonstrate how to use R to model differential equations, how to run code in parallel as well as creating high quality graphs. This project is implemented in Rmarkdown.

---