preamble
Many users who want to do quantification have been suffering from the lack of stable market sources, and I am one of them. But in fact, there are many free and easy-to-use market sources on the market. Here I recommend a few that I have used. I will give you a reference first
. Make a comparison:
| INSIGHT | Tushare | poly wide | Rice basket | |
|---|---|---|---|---|
| language support | java，python，c++,c# | python, r language, matlab | python | python |
| Market data level | millisecond-second | minute level | second level | second level |
| market data | K-lines of each granularity are free, tick data is free, and data is paid on a case-by-case basis | K-lines of each particle size need to be integrated | K-line payment for each granularity, and ticks are paid separately | K-line payment for each granularity, and ticks are paid separately |
| data file | complete | complete | complete | complete |
| registration process | The trial account can be opened immediately for free (2 months), and the official account is also free, but you need to become a customer of Huatai Brokerage | Recharge points to unlock, points need to spend money to buy, once a year | Annual Membership | Annual Membership |
| Financial Information Data | full, free | Generally, points are required | full, paid | full, paid |
| Types of Financial Information Data Provided | Industry classification, company data, financial data, derivative indicators (capital flow, ups and downs analysis, etc.), convertible bonds, ETFs, factor data (alphanet, barra), etc. | Industry classification, company data, financial data, dragon and tiger list, macro data | Industry classification, company data, financial data, dragon and tiger list, macro data, public opinion data, ETF, factor data | Industry classification, company data, financial data, dragon and tiger list, macro data, alternative data (search data, e-commerce data), ETF, factor data |
| Whether to support data localization | support | support | additional purchase required | additional purchase required |
| customer service response | quick | almost no customer service | quick | quick |
| Overview | 5 stars | 3 stars | 4 stars | 4 stars |
- Based on the principle of free and the most fragrant, INSIGHT data source is used as expected, and the response to data customers is very fast, and the documentation is relatively complete. The disadvantage is that the popularity is too low. In fact, the interface is very comfortable to use, and the degree of adaptation to the downstream is also very high.
1. INSIGHT data source (5 stars)
Many people have never heard of INSIGHT, which is actually a free market source released by Huatai. Huatai has always been very willing to invest in technology. Almost all kinds of data can be used for free. Except for the official account that requires Huatai customers, there is no other threshold, and the data quality is relatively high. Tushare will not often have wrong data. Case not repaired.
In fact, the data service is easy to use on the one hand, and on the other hand, it can solve problems in a timely manner. This is what INSIGHT does well, and technical support generally responds in real time. Generally speaking, the user experience has been good. Recently, I heard that real-time market announcements and news data may be accessed, combined with AI for labeling. This is still very important for me as an event-based strategy.
Attach the digital document link of the python version:
https://findata-insight.htsc.com:9151/insight_help/
Apply for a test account at:
https://findata-insight.htsc.com:9151/insight_help/sdk/SDKDownload/
The demo download is at:
https://findata-insight.htsc.com:9151/insight_help/python/BasedOnUsing/
2. Tushare data source (3 stars)
How should I put it, I actually love and hate this data source. I count it as the first batch of users. From the initial free to the current annual payment, I actually understand it quite well. As an open source data community, there is no It's easy, but now the price is getting higher and higher, but the service can't keep up. There are often errors in the data, and the customer service is basically unresponsive.
The data type is still very good, and the data classification is also very good, but at present, other data vendor documents are actually gradually catching up, and this advantage does not exist.
Overall, it’s good. After all, there are many sticky users. However, in the era of full name quantification, the data level is still at the minute level. It’s really not enough. I hope it will get better and better.
Attached link: https://tushare.pro/login
3. Jukuan + rice basket (4 stars)
In fact, these two have always been similar. The data types and functions provided by the two are basically similar. From the beginning, the quantitative community has gradually developed in the direction of 2B. In fact, there has always been a paradox. The strategy is to be a community. It doesn’t make sense. To be able to make money with a strategy, you need the concealment of the strategy. If there are too many people who know it, you won’t make money. Therefore, high-level strategies and investment are not reliable at all. Who has good things and will sell them on the community?
The quality of these two data is still good, but they are expensive, and if you need localized downloads, you need to spend extra money. The tick data is actually market public data. Huatai INSIGHT provides them for free, and I don’t know why they charge separately. The information data is quite complete, it is convenient to call, and the documentation is also good.
At present, I mainly use the local backtesting software backtrader. I dare not use the backtesting on the cloud side. These two companies actually have their own private placements, and you can’t find out if you steal the strategy. It's better to put everything locally for security
Attached link: https://www.joinquant.com/data