Explore Hot Concepts: New Energy Vehicles, Real Estate, Artificial Intelligence and the Data Feast of 5G

After Thematic Big Data launched its service to the public, we provide the top 20 hot topics every hour during the daily trading period. Taking stock of the past 5 years, the top hot themes are New Energy Vehicles, 5G, Pharmaceuticals, Real Estate, Food, and Artificial Intelligence.

Thematic big data can be used for index enhancement, constructing thematic rotation strategies, reversal factors and other multi-application scenarios; here is a brief chat about thematic big data.

1）Statistics on the number of topics

The inclusion of the theme master list started in 2017 with a machine+manual method of refining themes, and there are currently 1,426 active themes.

2）Number of thematic constituents

Many themes, large and small, exist in the market, like a huge network. Large themes have more than 600 constituent stocks, while smaller themes may have less than 10 constituent stocks. Below is a count of the number of constituent stocks of the themes.

The data from the Theme and Stock Correlation Table for the day 20240630 was analyzed. There were a total of 1,325 active themes on that day.

The vast majority of thematic constituents are within 300 of each other.

We set several ranges for the number of constituents and count the number of themes within each range. The chart above shows that most of the themes have a number of constituents within 10-150. A theme with too large a number of constituents will not provide a precise enough description of the common fundamental drivers; a theme with too small a number of constituents will have a weak impact on the market and will not be worth paying attention to. Themes can therefore be filtered according to the granularity required for research, and it is suggested that one can focus on themes within the 30-150 range.

3）Attempts at factor effects for themes

Themes, in fact, are a way of clustering individual stocks based on a certain characteristic. This characteristic, which may be derived from industry, policy, or events, represents a trading sentiment shared by the market. Theme index, can be regarded as the industry, policy, events, these concepts of the materialization of the underlying.

As of today, there are 1,426 valid themes available in the Theme Big Data, with most themes having between 10-150 constituents. A small number of themes have overlapping and inclusion relationships and can be filtered appropriately.

When constructing the theme rotation strategy, we found that there are some alpha effects of themes, such as cross-sectional momentum effect, popular theme alpha effect, and emerging theme alpha effect. Hot themes, emerging themes, provide an idea of filtering the theme pool, which derives part of the return from alternative information at the theme level. Similarly, with reference to the multi-factor model of individual stocks, we can also construct theme factors from the bottom up and screen theme pools based on thematic factors, which is another perspective of return sources. In this paper, we try to construct 2 theme factors from momentum and value perspectives, and synthesize them for theme screening to get a theme rotation strategy with an annualized return of 10.69% and an information ratio of 0.88. Combining the two perspectives to construct a high quality and popular theme rotation strategy, the annualized return reaches 22.61% and the information ratio reaches 2.24.

Finally, we try to compute individual stock factors using theme information to compute individual stock exposure to quality themes. We try to calculate the rising theme exposure factor, and the ten-group long-short portfolio on ZZ800 reaches an annualized return of 7.49% with an information ratio of 1.3.

If you are interested in test research on the topic, feel free to contact datayes for more in-depth information!