Dylan Poh Portfolio

Linkedin - https://www.linkedin.com/in/dylan-poh/

Project 1: Covid19 Dashboard Web App using Python (Plotly Dash)

Feature on Plotly - https://moderndata.plotly.com/covid-19-dash-apps/

Read more on https://www.reddit.com/r/Python/comments/g5ymoy/six_months_into_python_and_data_science_my_first/

Create your own Covid19 Data Visualization dashboard web app with free resources using python only.

• Engineered, clean and extract features from Johns Hopkins University Center Dataset to visualize meaningful data.
• Built a client facing API using Plotly Dash

Version 1.0 Dashboard
Click here to access the web app

Version 2.0 Dashboard based on bootstrap 4
Click here to take a look at the web app

---

Project 2: Image Story Generator - Collaboration Project

A webapp created with FLASK API that utilizes computer vision and natural language processing to generate dramatic short stories using features from images.

 

AI Apprenticeship Programme (AIAP®) Capstone Project- Aug 2020

---

Project 3: FASTAI Image Classification with Azure Function and Telegram bot Deployment

A mini project that uses FASTAI Image Classification to classify 3 types of dogs namely, Pug, French Bulldog, and Boston Terrier. The trained model will be deployed to Azure Function and the user can send in their dog images to Telegram bot and receive the breed type.

Telegram Bot Demo( Welcome to try at Dog Whisperer Bot - t.me/DogWhispererAi_bot)

---

Project 4: Web Scraping Angellist Companies

• Web scraping Angellist companies and Data Visualization with Selenium Webdriver
• Vsualization Jupyter Notebook - Angellist_Data_visualization.ipynb in folder Data_Visualization

AngellistTop 20 Market Singapore

---

Project 5: Housing Prices Competition

XGboost with hyperopt (Mean Absolute Error: 14876) Rank 627th on Kaggle - Top 4% Code uploaded in the folder XGboost_Hyperopt. ***

---

Project 6: Occulus VR using Unity - Music Slime Beat Saber Clone

A clone Oculus VR Beat Saber Game for my personal project. (Title: Slime Music Beat)

My learning experience of Unity started 4 months ago, and by gathering free tools online, I am able to create this game within one day. Listed below are the youtube videos I followed to create the majority of the game. The rest of the scripts, animations and etc are based on my 4 months learning experience to glue them together.

You can view the gameplay here https://www.youtube.com/watch?v=eiCgn92ClfY&t=1s

---

Project 7: Stock_analysis_Pandas_Scikit_Learn

Using Pandas and Scikit Learn to pull data from Yahoo Stock for analysis

This script is based on https://towardsdatascience.com/in-12-minutes-stocks-analysis-with-pandas-and-scikit-learn-a8d8a7b50ee7

The Quadratic Regression 2 with the highest confidence among the 4 regression models was choosen: 0.7919698975806313 Blue line shows the forecast based on Quadratic Regression 2, indicating a downward trend of the stock.