Experiences 經歷

  • Data Analyst 資料分析師
    New Power Party 時代力量
    Taipei 台北市

    Responsible for various data-driven projects:

    AI Implementation:

    • Designing and deploying AI solutions
    • Created cost-effective AI tools for internal use
    • Integrated with OpenAI and Hugging Face APIs to build RAG model and other AI tools

    Financial Process Improvement:

    • Built automation tools for financial approval system and accounting workflows
    • Developed a web app using Django, MySQL

    News Sentiment Analysis:

    • Tracking media coverage and public opinion
    • Implemented web scrapers to collect data from major news sites
    • Created interactive dashboards for real-time sentiment tracking

    Social Media Metrics:

    • Developed a interactive dashboard using Streamlit, Meta Graph API, Youtube API, Notion API
    • Provided weekly reports and monthly competitive product analysis
    • Article: Social Media Analysis Fundation

    System Integration:

    • Developed CRM system and file management system
    • Reviewed and restructured data access permissions and policies

    AI 導入專案

    • 設計及開發 AI 工具使用 OpenAI API、Hugging Face 開源工具
      建構 RAG 生成式檢索、語音轉文字等工具
      提供符合成本效益的 AI 解決方案

    財務流程改善

    • 開發財務簽核系統Web App 全端開發，採用 Django 架構
      MySQL 資料庫、串接 Google API 及 GCP 架設
      規劃及設計專案預算追蹤、視覺化
      自動化財務簽核流程
    • 建立財務流程自動化工具金流平台資料與內部系統自動化對帳
      使用 Google App Script 建立收據寄發自動化

    新聞輿情分析

    • 設計及開發輿情監測工具執行網頁爬蟲蒐集主流新聞媒體及論壇內容
      使用 matplotlib、streamlit 建立可互動視覺化圖表
      採用自動化工具定時更新，整合多元資料來源

    社群指標

    • 串接 Meta Graph API、Youtube API 等取得 FB、IG、YT 等粉專社群資料
      利用 Notion API 同步更新最新資料並提供各項指標變化
      專案經驗分享：社群分析基礎

    系統整合

    • 整合各部門需求，規劃 CRM 資料庫、檔案管理規則等
    • 盤點系統、資料庫的使用者權限，提升系統安全及權責透明
  • Research Assistant 研究助理
    Center for Behavioral and Data Science 行為與資料科學研究中心
    Taipei 台北市

    Worked on several large-scale data analysis projects:

    College Entrance Examination Center

    • Identified correlations between student backgrounds and test results
    • Built data visualizations
    • Provided policy-making reference

    Fiscal Information Agency

    • Analyzed and processed tax and economic data from Fiscal Information Agency and other departments
    • Applied statistical methods to identify relationships between income levels and health risks

    Policy Impact Analysis

    • Developed quantitative indicators for policy evaluation
    • Provided evidence-based recommendations to support long-term policy

    執行多項巨量資料分析專案：

    大學入學考試中心

    • 分析成績與背景因素之相關性
    • 產出資料視覺化
    • 提供教育政策制定參考

    財稅資料中心

    • 分析稅務母體資料串接各部會資料如內政部、教育部等
    • 建立統計模型探討所得水準與健康風險之相關性

    政策影響評估

    • 建立量化指標提供政策制定之科學依據
    • 為教育、財稅及健康政策領域提供實證資料分析依據

Education 學歷

  • MSc in Social and Economic Data Science 社會與經濟資料科學碩士
    University of Konstanz 康斯坦茲大學（Konstanz University）
    Konstanz, Germany 德國康斯坦茲

    Main courses:

    • Data Mining
    • Data Visualization
    • Database
    • Big Data Management & Analysis
    • Probabilistic Modelling for Computer Scientists
    • Applied Time Series Analysis
    • Machine Learning
    • Research Design
    • Network Modeling and Relational Machine Learning

    重點課程：

    • 資料探勘
    • 資料視覺化
    • 資料庫
    • 大數據管理與分析
    • 電腦科學機率模型
    • 應用時間序列分析
    • 機器學習
    • 研究設計
    • 網路模型與關聯機器學習
  • Bachelor of Economics 經濟學學士
    National Taiwan University 國立台灣大學
    Taipei 台北市

    Main courses:

    • Statistics
    • Microeconomics Theory
    • Macroeconomics Theory
    • Econometrics Theory
    • Industrial Organization and Firm Strategy
    • Labor Economics
    • Neuroeconomics
    • Market and Economic Development
    • Quantitative Methods

    重點課程：

    • 統計學
    • 個體經濟學理論
    • 總體經濟學理論
    • 計量經濟學理論
    • 產業組織與廠商策略
    • 勞動經濟學
    • 神經經濟學
    • 經濟發展
    • 量化方法
  • Bachelor of Business Administration 企業管理學士
    National Cheng Kung University 國立成功大學
    Tainan 台南市

    Main courses:

    • Business Statistics and Data Analysis
    • Theory and Strategy of Financial Management
    • Accounting
    • Marketing Research
    • Management Information Systems
    • Psychology
    • Risk Management
    • Behavioral Science

    重點課程：

    • 商管統計資料分析
    • 財務管理理論與策略
    • 會計學
    • 行銷研究
    • 資訊系統管理
    • 心理學
    • 風險管理
    • 行為科學

Skills 技能

Languages 語言

Chinese (Native), English (Proficient), German (A2), Japanese (N4) 中文（母語）、英文（精通）、德文（A2）、日文（N4）

Programming Languages 程式語言

Python, R, SQL, JavaScript, HTML, CSS, MySQL Python、R、SQL、JavaScript、HTML、CSS、MySQL

Data Analysis & Visualization Tools 資料分析與視覺化工具

R, Stata, Tableau, Matplotlib, Plotly, Seaborn, Streamlit R、Stata、Tableau、Matplotlib、Plotly、Seaborn、Streamlit

Machine Learning & Deep Learning 機器學習與深度學習

TensorFlow, Scikit-learn, Pandas, NumPy TensorFlow、Scikit-learn、Pandas、NumPy

Full-Stack Development (Web App) 系統全端開發（Web App）

Django, Docker, GitHub Django、Docker、GitHub

Web Scraping & Data Collection 網頁爬蟲與資料蒐集

Selenium, BeautifulSoup, Meta Graph API, Google API Selenium、BeautifulSoup、Meta Graph API、Google API

Publications 發表

  • Economic Forecasting: A Comparison of Time Series Methods and Neural Network 經濟預測：時間序列方法與神經網絡模型比較

    Authors: 作者: Hui-Chih Shen 沈暉智

    Master Thesis, University of Konstanz 康斯坦茲大學碩士論文 • 2023

    This paper explores the forecasting performance for global shipping rates, comparing time series models (VAR) and neural network models for predicting the Baltic Dry Index (BDI) and Kilian Index. The research finds that VAR models perform better in most scenarios, while neural networks are suitable for long-term trend forecasting. Additionally, the impact of major international events significantly affects forecasting accuracy; when these events are excluded, the performance difference between different models diminishes. 探討國際乾貨航運價格預測，比較時間序列模型（VAR）與神經網絡模型（RNN/LSTM）的預測表現。研究發現，VAR 模型能勝任大多數情境的預測，而神經網絡模型適合長期趨勢預測。此外，重大國際事件對預測準確性有顯著影響；當排除這些事件時，不同模型之間的表現差異會減小。

  • The Impact of Household Income and Assets on Children's Education - Evidence from Tax Data of 1993-1995 Birth Cohorts and Their Parents 論家戶所得與資產對子女教育之影響 — 以1993–1995 出生世代及其父母稅務資料為例

    Authors: 作者: Hui-Chih Shen, Ming-Jen Lin 沈暉智、林明仁

    Taiwan Economic Review 經濟論文叢刊 • 2019

    This paper explores intergenerational income mobility through the education channel. We use 2000–2015 household income data provided by the Fiscal Information Agency. We find a strong correlation between household income and access to higher education. The median of yearly household income in the full sample is 0.86 million NTD, while the medians among National Taiwan University students and public university students are higher than 1.5 and 1 million NTD respectively. We look into asset distribution in every education category, finding a huge amount of wealth inequality and rigidity of intergenerational mobility. The positive correlation between education subsidy and students' household income shows that government subsidy may lead to income redistribution from the poor to the rich. 分析財稅資料中心提供 2000-2015 年之母體稅務資料，探討親代所得與資產如何影響子代教育取得之機會。為碩士論文改寫刊登於經濟論文叢刊、並與報導者合作發表專欄

  • Can Education Transform Lives? The Reverse Redistribution in Higher Education 【數字說話】教育能翻轉人生嗎──高教的反向重分配現象

    Authors: 作者: RedHerrings, Hui-Chih Shen, The Reporter Editorial Team RedHerrings、沈暉智、報導者團隊

    The Reporter 報導者 • 2018

    An analysis of educational inequality and socioeconomic mobility in Taiwan's higher education system. 探討台灣高等教育系統中的教育不平等與社會經濟流動性。