About
I'm Cheng-Ching Lin, a data scientist in Taiwan. I am passionate about data analysis, causal inference, XAI, and AB testing. After graduation, I always think it is lonely to study these things on my own. Therefore, I create this personal website, and I want to share with you all I study about.
Due to my inherent laziness (yes, I'm guilty!), I have been utilizing chatGPT to generate the initial drafts of my articles. But fear not, I take the responsibility of verifying the accuracy and ensuring the conveyance of my intended message before publishing.
Experience
Data Scientist
Vizuro (a startup solution provider of causal AI), Taipei, Taiwan – (Mar 2023 - Present)
- Improved the existing causal discovery algorithm by incorporating a diffusion model, resulting in a more accurate identification of distribution and importance of causal relationships.
- Improved the causal discovery algorithm to support continuous and discrete values, as well as big data with over 50k features, resulting in a more robust and scalable solution for analyzing complex datasets.
- Implemented new features that allow customers to visualize response curves, enhancing their understanding of causal relationships and enabling more informed decision-making.
Data Analyst Intern
Appier, Taipei, Taiwan – (Jun 2022 - Dec 2022)
- Analyzed customer user journey in e-books websites, and suggested a campaign that helps them to increase registration rate over 2 times. The report becomes the template of regular yearly reports for customers.
- Established a key metric to classify customers in e-book websites, and stimulated re-think about the free-book strategy.
- Identified the main driver behind consumers' motivation to purchase a new smartphone, resulting in at least a 30% increase in new smartphone purchases after implementing my recommendation.
Group Data & Analytics Office – Summer Intern
China Development Financial, Taipei, Taiwan – (Jul 2021 - Aug 2021)
- Targeted 1% of potential customers who can be cross-sold in KGIS, such that at the same cost, the number of successes made by our model is at least 5 times that of the original strategy. This model process still works in CDF.
Project
Predicting revenue-maximizing promotions for Shopline merchants
Shopline
- Offered prediction for each promotion with Random Forest so that merchants can decide and expect the revenue before the promotion begins.
- Provide merchants with the optimal revenue promotion campaign settings. The simulation results show that the revenue can be increased 10 times by using optimal settings.
Predict Risk Insurance Case
Nan Shan Life
- Predict risk cases using LGBM with semi-supervised learning, and the top 10% lift 2 times than baseline.
Analysis Reflow Oven in SMT production line
Foxconn
- Analyzed the relation among oven temperature, the concentration of certain gases, and the quality of PCB board, and constructed a guideline to tune the setting of oven temperature and gases.
Education
Master's Degree in Statistics
2021 - 2023
National Tsing Hua University
Exchange Student in Physics
2017 - 2018
Technion - Israel Institute of Technology
Bachelor's Degree in Physics
2014 - 2018
National Tsing Hua University
Additional Skills
- TOEFL ibt test: 96/120
- TOEIC: 865 (Golden Certificate)
- GRE test: 318/340 (Verbal 150, Quantitative 168, Writing 3.5)
- Technical Skills: RapidMiner, R, Python, SQL, Tableau, Excel