Introduction
Financial data is the basis of quantification. Without data quantification, it is impossible to start. As competition in the industry intensifies, quantification has higher requirements for high quality—faster, more complete, and more accurate data. After all, quantification cannot be lost at the starting line. Nowadays, there are various quantitative data sources. I found that many friends are not very clear about this. I have also spent a lot of time on it. This time I will share some known data. Interested friends can follow the official account and join the group. comminicate.
Quantitative data source comparison
Quantitative data sources are divided into the following types: open source quantitative data, data sources provided by securities firms/quantitative trading platforms, professional data service companies, and self-grabbing and cleaning methods.
Open source quantitative data
By grabbing various financial websites or public financial data, cleaning, processing, storing and then opening them up, we can provide financial data needs for quantitative learners.
-
BaoStock - a free, open source securities data platform (no registration required), which obtains securities data information through python API. The returned data format is pandas DataFrame type. It also supports BaoStock's data storage function to save all data locally for analysis. . The advantage is that it is free, but the disadvantage is that the data is incomplete
-
tushare – divided into tushare and tushare pro. tushare pro data has a wide coverage, but there is a point limit when the number of acquisitions is large. The old version of the API only provides basic daily data. The tushare pro cash points recharge ratio is 1:10. For example, if you recharge 200, you will get 2,000 points. The points are valid for one year, and the data points will not be reduced if you use it. 2,000 A-share points can be used, but the frequency is limited, with a limit of 200 requests per minute and a limit of 100,000 api requests per day. Daily trading of Hong Kong and US stocks requires at least 5,000 points. Minute-level data requires separate permissions. It has nothing to do with platform points and needs to be applied for separately.
-
akshare - a Python-based financial data interface library, which aims to realize fundamental data, real-time and historical market data, and derivative data from financial products such as stocks, futures, options, funds, foreign exchange, bonds, indices, and cryptocurrencies. A set of tools from data cleaning to data implementation. The akshare api interface changes frequently and the data format is not universal.
-
yfinace - Yahoo financial data acquisition, you need to use a proxy to access
-
easyquotation--python obtains Sina/Tencent's full market conditions in real time, but historical data cannot be obtained
-
efinance – Free open source Python library for obtaining stock, fund, futures, and bond data
Brokerage/quantitative trading platform
-
Jukuan Data JQData – Jukuan provides local quantitative financial data service jqdatasdk. You can apply for a three-month trial period, and a mobile phone number can only be registered once.
-
rqdata--data service provided by Mikuang. Free trial for 15 days, trial account has a daily quota limit of 50MB
-
tqsdk – The free version of TqSdk provides all futures, commodities/financial options and real-time quotes of SSE 50, CSI 300 and CSI 500. The professional version of TqSdk provides real-time and historical quotes of A-share stocks.
-
futu openapi: The quantitative API provided by Futu can obtain historical and real-time market prices. Different data acquisition permissions are corresponding to the account amount level.
Professional data services company
-
Wind Data Service data service - about 30,000 to 60,000 yuan/year, you need to consult sales for details, domestic financial data Wind is relatively more complete
-
ifind — Flush financial data terminal, priced at about 1/3 of wind
-
choice data - Oriental Fortune data terminal, the quality of choice is said to be not as good as ifind
-
Bloomberg data service - about $20,000/year, please consult sales for details
Capture & clean data yourself
Friends with good programming skills and time can also capture and clean the data by themselves. The advantage is that the data quality is guaranteed and can be processed according to your own requirements; the disadvantage is that it requires certain programming skills and is more time-consuming and labor-intensive.
What is more common is that when the above data sources cannot cover specific data, you can crawl and supplement it yourself.
Conclusion & Communication
Follow the official account for more content. At the same time, you can also get invitations to join the community, communicate and discuss with many quantitative practitioners and enthusiasts, and not miss the latest industry development and technological progress.
Public account: Zhugeshuotalk
Writing articles is not easy. If you think this article is helpful to you, please give it a like.
reference
-
https://github.com/rchardzhu/awesome-quant-cn
-
https://tushare.pro/document/1?doc_id=13
-
http://baostock.com/baostock/index.php/Homepage
-
https://github.com/akfamily/akshare
-
https://www.joinquant.com/data
-
https://www.zhihu.com/question/20373441
-
https://www.zhihu.com/question/268703873
-
https://www.zhihu.com/question/39264388