Applying NLP tools to earnings calls, company filings and financial documents is not new to the world of alternative data. Drawing specifically on earnings calls, many will also be familiar with post-earnings call announcement drift (PEAD). In this report, we review literature that discusses how PEAD can be improved using linguistic approaches such as NLP.

We summarise a 2021 study that analysed J-REIT data. The study evaluates the effects over time of three price-forming factors on the capitalisation rate of office real estate in Tokyo’s business districts.

We summarise a study examining how the sentiment of Chinese retail investors can be used to predict stock price movement. The study found alternative data from Xueqiu, a Chinese social media platform for retail investors, improved the accuracy of Chinese A-share price predictions.

We summarise key findings from a research paper that examines using sentiment analysis of news sources as a leading indicator of economic performance in Japan.

We summarise the key findings from a recent research paper that examines the dynamic relationship and volatility spillover effect between the Chinese stock market and Chinese commodity markets.

Given some of the limitations of in-situ biodiversity data, the authors explore using satellite data to derive biodiversity-related physiological proxies. We summarise the data sources, advantages and limitations discussed in the research, and highlight potentially relevant satellite data providers.

We summarise a 2022 paper that examines whether Bitcoin and the aggregate cryptocurrency market acts as a diversifier, hedge or safe haven to US stocks, bonds, US dollar, commodities, real estate and gold.

Published research suggests the securities lending market is opaque and subject to non-competitive features. We summarise the findings of this research and examine the suitability of IHS Markit’s Securities Finance product in assessing securities lending market activity.

We summarise a recent paper based on a high-frequency trading firm that adopts deep neural network (DNN) machine learning systems for trading functions. The paper considers how we can begin to explain the decisions of DNNs, despite their opacity.

With current concerns over rising inflation across major economies, we summarise two studies that anticipate official CPI figures ahead of their release. These studies use high-frequency alternative data sources, which are found to improve the forecast accuracy of inflation models and provide timely insights into global pricing dynamics.