• Email Address: [email protected]
• LinkedIn :

1. Quantitative Analysis Consultant Working Student
   • Company: RIVACON/Energy Company
   • Period: 09/2023-04/2024
   • Brief Job Description: Intraday and Day-ahead energy market price and wind infeed analysis
2. Full-Time Assistant Researcher
   • Institution: National Taiwan University/Taiwan Academia Sinica
   • Period: 02/2021-06/2021
   • Brief Job Description: Long-term Taiwan food demand/supply analysis and prediction. Also, the literature review on extreme weather scenarios
3. Marketing Intern
   • Company: Carousell
   • Period: 04/2018-06/2018
   • Brief Job Description: Questionnaire design and analysis

This project replicates the paper Stefano DellaVigna and Devin Pope, 2018, "What Motivates Effort? Evidence and Expert Forecasts", Review of Economic Studies, 85(2): 1029–1069. The objective of this project is to rewrite all the codes in a more effective way and re-organize them. The main contribution of the paper is to quantitatively bridge the rewards and people's efforts by different functions; based on different reward mechanisms setup, people make different levels of effort. The details can be found in the following repository.

• EPP_Final

This is the repo for my master thesis, titled "Applying Generalized Random Forest to Analyzing the Heterogeneous Effects in Regression Discontinuity Design". It can be separated into four main sections. First, mathematics definitions, assumptions and proof. Second, typical Monte-Carlo Simulation with general and point estimation cases. Third, applying the WGAN to generate data similar to real-world data in terms of distribution, mean, and variance. In the end, the generated data are again applied in Monte-Carlo Simulation. Fourth, using RDD on GRF in the real-world dataset and comparing with the published research result. The details can be seen in the following repository.

• Applying Generalized Random Forest to Analyzing the Heterogeneous Effects in Regression Discontinuity Design

This project is the expansion of the paper "Comparison of Single and Ensemble Classifiers of Support Vector Machine and Classification Tree". Journal of Mathematical Sciences and Applications, 2(2), 17-20 (Utami, I.T., et al., 2014). In this project, the classification problem is the main focus; however, instead of applying SVM, several tree-based methods are discussed, and bagging and boosting methods are also included. Furthermore, both large and small dataset scenarios are discussed in the project. ROC curve, confusion matrix, and misspecification rate are taken as evaluation criteria. The details can be found in the following repository.

• The analysis and comparison of the tree-based method with different simulation assumptions

This project is aiming at conducting a comprehensive analysis of Disney Plus from social media perspectives. The analysis has been done in the following sub-sectors: Real-time reaction according to the recent release of the Movie or series, a sentiment analysis will be conducted based on the text from Twitter, YouTube, and Reddit. On top of that, we will analyze the impact of influencers' opinions on the platform and offer feasible improvements to the platform. This project requires natural language processing skills(NLP) and Web-scrawling skills. The full version of the analysis with the codes can be found in the following repository

• Social Media Analysis of Disney Plus

This is a small exercise that is aimed at discussing the pros and cons when comparing the linear regression models and regression tree models. In this exercise, the DGP which is suitable for linear regression models and regression tree was explicitly mentioned. The details can be found in the following repository.

• Data Generating Process and Simulation Exercise

This is an exercise from the effective programming courses, and the exercise is based on the paper estimating the Technology of Cognitive and Noncognitive Skill Formation by Cunha, Heckman, and Schennach (CHS), Econometrica, 2010. The objective of the task is to manage the data in an elegant, effortless way by different function setups. In a large, dirty dataset(many non-informative or duplicate data), the task includes, filtering data according to the requirement, transposing, turning characteristical data into numeric data, and turning some data information into columns. Furthermore, automatically plotting several specified regression and correlation graphs is also required. The details of the task and the solution can be seen in the following repository.

• Data-Management-And-Ploting

This report is aiming at using different machine learning skills and algorithms on Blablacar's real-world data to do a data-driven marketing analysis. In this report, Random Forest, Decision Tree, Adaboost, SVM, Voting method and Neural network have been used for pricing model and be compared. Besides, k-means method and linear regression were applied for market analysis and causal inference. The details can be found in the following repository.

• Blablacar_Analysis