Wai Kien

Biomedical Science background // Data Enthusiast

Hola, I'm Wai Kien.
I like learning, data and learning with data.

I'm currently working as a Data Analyst. To put things into perspective, the data industry is mainly classifed into 3 spectrums:
Data Engineer ⇄ Data Analyst ⇄ Data Scientist.

Data Scientist

I am eager to explore all 3 and ultimately specialise in one of them 👨🏻‍💻

Work Experience

2023 - present

mecp2
Shopee!

Business Intelligence - ShopeePay & SeaMoney

Conducting data-driven business iterations for Fintech

Python • SQL • Spark • Business Analytics • Statistics

2021 - 2022

mecp2
In Layman terms: A. (Left) MeCP2 stops growth by binding to DNA and recruit "helpers". (Right) MeCP2 binds to DNA and stops growth on its own. B. We add MeCP2 into controlled bacteria without "helpers", and see if it stops growth.

Research Placement - MRC London Institute of Medical Sciences

Investigating epigenetics mechanisms

I joined the lab of Prof. Peter Sarkies that focuses on epigenetics and evolution.

During this one year, I investigated this protein called MeCP2. To give some context, DNA methylation adds chemical flags to some bases in the DNA strand. In human cells, methylation turns gene off. It is known that MeCP2 helps methylation to do its job by binding to methylated DNA strands and recruit "helper" proteins. Loss of MeCP2 leads to a serious genetic brain disorder known as Rett syndrome, but scientist still don't understand what MeCP2 does and why not having it leads to such as severe disease.

To study this, we made a simple model in bacteria E. coli. Using this we discvoered that MeCP2 can affect the growth of E. coli on its own without other "helper" proteins. We found out this is because MeCP2 directly interferes with the growth mechanisms of E. coli by incorrectly turning on and off some genes in E. coli.

MSExcel • R (DESeq2, clusterProfiler, AnnotationDbi, seqinr, ggplot2) • Differential Expression Analysis • Gene Ontology Analysis • Statistics


Projects

2023

train

How My iPhone and Mac Usage Affects My Daily Life

I wrote an article about my analysis on my personal iPhone and Mac data.

Here's some interesting findings:

  • Step count ↑ = Screen time ↓
  • I tend to be more productive early-week as my screen time isn't sky high
  • I pick up my phone more than 100 times per day
  • I spend >10 hours on my phone and laptop, per day, wow

Read The Full Article Here

Python • Data Visualization

2023


bitcoin

LSTM Bitcoin Prediction

Building a Long Short-Term Memory (LSTM) neural network model that predicts $BTC close price, this is also known as multivariate time series forecasting.

Workflow wise, I fed the following data into the model and ask it to predict future price:

  • Historical $BTC Price and Volume
  • How many people are holding >1000 $BTC
  • $BTC mining hashrate and difficulty
  • Number of active $BTC addresses
  • Transaction Counts

I then implemented a simple trading simulation to backtest the model. It outperformed "random long and short" strategy by ~8% and "buy and hold" strategy by ~300%.

Python • pandas • numpy • TensorFlow • Keras

Languages

Python R SQL HTML CSS JS

Databases

spark postgre mongodb trino

Socials

linkedin medium github gmail

Miscellaneous

Something I'm proud of

TypeRacer.com scorecard for user feikien99

Something I'm not proud of

Leetcode Stats