近年來金融科技(FinTech)在全球蓬勃發展，使用最新的科技結合過去的金融服務，讓企業提供更有效率的服務，也讓用戶有更便利即時的使用體驗。而其中最常見的應用是利用蒐集用戶或領域的資料進行數據分析，例如各家金控都大力發展的「理財機器人」便是利用企業背後龐大的資料庫數據結合使用者的資訊進行分析，提供用戶最佳的投資組合及策略建議。此類應用是基於從大量數據中找出相關性和隱藏在其中的關鍵因子，可想而知數據本身的完整性和正確性對於分析結果會有舉足輕重的影響，如何取得及處理大量原始數據(raw data)是數據分析的首要挑戰。

今年(2020)三月台灣證券市場逐筆交易正式上線，對於短線波動帶來不小影響，有些券商系統無法處理大量資料而造成投資人下單與看盤上諸多問題。如何建構一套量化交易系統，能夠處理更高頻的逐筆交易進行分析，便是本論文將探討的主題。本系統蒐集臺灣證券與期貨選擇權市場數據，包含逐筆交易成交明細、委買賣掛單揭示資訊、一分K線、三大法人買賣超、大額交易人交易部位等數據，經過多方交叉比對及處理後確保資料的正確性，建構當代RESTful與WebSocket架構之量化交易分析平台，供市場交易研究與量化策略開發回測使用，以促進投資人理性分析與增進投資效益首要目標。

本論文所開發之API及視覺化平台，將對金融科技發展帶來極大的益處，解決目前金融交易難以視覺化量化表示及驗證等問題外，研究人員可專心在設計演算法與驗證有效性，交易員可以獲得即時直觀的分析圖表，而不須花費心力在資料處理。對於推廣以資料為根本的量化交易(Quantitative Trading)投資策略上也能更加順利，讓臺灣的金融交易制度以及投資人的思維可以跟上全球先進國家。

In recent years, FinTech has flourished around the world, using the latest technology combined with past financial services, allowing companies to provide more efficient services, and also to make users more convenient and instant experience. The most common application is to collect data from users or fields for data analysis. For example, the "financial robots" developed by various banks are using the huge database data behind the enterprise to analyze the information of users and provide users. The best portfolio and strategic advice. Such applications are based on finding correlations and key factors hidden in a large amount of data. It is conceivable that the integrity and correctness of the data itself will have a significant impact on the analysis results. How to obtain and process the raw data (raw Data) is the primary challenge of data analysis.

In this paper we collect data of Taiwan's securities and futures market. In addition to basic volume and price data, it will integrate information such as chips, brokerage points, financial reports, internal and external trading, margin financing and securities lending, and ensure data after cross-matching and processing. The correctness is to construct an API platform that conforms to the RESTful architecture for academic research institutions to apply for, in order to promote the overall research progress as the primary goal.

The API platform developed by the project will bring great benefits to the development of financial technology. In addition to solving the problem that financial transaction data is difficult to obtain and verify, researchers can concentrate on designing the structure of the analysis data without spending effort on data processing. The investment strategy for promoting data-based quantitative trading can also be smoother, allowing Taiwan's financial trading system and investors' thinking to keep up with the advanced countries of the world.

[1]Candlestick chart.https://en.wikipedia.org/wiki/Candlestick_chart.[2]Raymond james 41st annual institutional investors conference.http://investor.cmegroup.com/static-files/47110b0f-6398-4fe4-8578-a804402b954e.[3]Kimon P.Valavanisb George S.Atsalakisa. Surveying stock market forecastingtechniques–part ii: Soft computing methods.Expert Systems withApplications, 36:5932–5941, April 2009.[4]Yi-Ling Lai. Using reinforcement learning to establish taiwan stock indexfuture intra-day trading strategies.National Taiwan University, July 2009.[5]Greg Kemnitz M R Stonebraker. The postgres next generation databasemanagement system.Communications of the ACM, October 1991.[6]Robert Kissell Morton Glantz.Multi-Asset Risk Modeling: Techniques for aGlobal Economy in an Electronic and Algorithmic Trading Era. AcademicPress, 2013.[7]American ohlc chart.https://zh.wikipedia.org/wiki/%E7%BE%8E%E5%9C%8B%E7%B7%9A.[8]Rekhit Pachanekar. Automated trading systems: Architecture, protocols,types of latency.https://blog.quantinsti.com/automated-trading-system/, 2019.32
[9]Brian G. Peterson. Developing backtesting systematic trading strategies.https://www.researchgate.net/publication/319298448_Developing_Backtesting_Systematic_Trading_Strategies, 2017.[10]Rabbitmq topics.https://www.rabbitmq.com/tutorials/tutorial-five-python.html.[11]I-Yuan Tang. The empirical research of the taiwan stock exchangecapitalization weighted stock index bid/ask volumes variation operationstrategy.National Kaohsiung University of Science and Technology, June2011