Financial Sentiment Analysis (FSA) is a critical application of sentiment analysis in the financial domain, focusing on investor sentiment and financial textual sentiment. It has gained significant attention in recent years due to the complexity of financial markets and the need for automated tools to support financial decision-making and forecasting. FSA research is divided into two main streams: one focused on developing techniques and improving performance through human-annotated datasets, and the other on applying financial sentiment for downstream applications in financial markets, such as hypothesis testing and predictive modeling. This survey provides a comprehensive review of FSA research, covering both technical and application aspects, and proposes frameworks to understand their interactive relationship. It defines the scope of FSA studies and conceptualizes the relationship between FSA, investor sentiment, and market sentiment. The survey also summarizes major findings, challenges, and future research directions for both FSA techniques and applications. FSA involves analyzing financial texts to detect sentiment, which can be explicit or implicit. Financial sentiment is categorized into market-derived and human-annotated sentiments. Market-derived sentiments are derived from market dynamics, while human-annotated sentiments are labeled by professionals or investors. FSA has become a prominent research topic due to the increase in online financial data. Recent FSA research has shifted from human-annotated to market-derived sentiment. The survey reviews the latest FSA studies, offering a dual perspective from both technical and applied standpoints. It also connects FSA with other disciplines such as information systems and finance, reviewing the principles of financial forecasting and supporting the market predictability of financial sentiment from a financial theory perspective. The survey discusses the relationship among market sentiment, investor sentiment, and financial textual sentiment, highlighting their interdependencies. It also reviews the scope of FSA studies, the challenges in FSA techniques and applications, and the interactive relationship between FSA techniques and applications. The survey identifies six areas that cause FSA failure, including irrealism mood, rhetoric, dependent opinion, unspecified aspects, unrecognized words, and external reference. It also discusses the evaluation metrics used in FSA, including regression and classification metrics, and the methods used in FSA, such as lexicon approaches, machine learning approaches, and deep learning approaches. The survey concludes that FSA is a critical area of research with significant implications for financial markets and decision-making.Financial Sentiment Analysis (FSA) is a critical application of sentiment analysis in the financial domain, focusing on investor sentiment and financial textual sentiment. It has gained significant attention in recent years due to the complexity of financial markets and the need for automated tools to support financial decision-making and forecasting. FSA research is divided into two main streams: one focused on developing techniques and improving performance through human-annotated datasets, and the other on applying financial sentiment for downstream applications in financial markets, such as hypothesis testing and predictive modeling. This survey provides a comprehensive review of FSA research, covering both technical and application aspects, and proposes frameworks to understand their interactive relationship. It defines the scope of FSA studies and conceptualizes the relationship between FSA, investor sentiment, and market sentiment. The survey also summarizes major findings, challenges, and future research directions for both FSA techniques and applications. FSA involves analyzing financial texts to detect sentiment, which can be explicit or implicit. Financial sentiment is categorized into market-derived and human-annotated sentiments. Market-derived sentiments are derived from market dynamics, while human-annotated sentiments are labeled by professionals or investors. FSA has become a prominent research topic due to the increase in online financial data. Recent FSA research has shifted from human-annotated to market-derived sentiment. The survey reviews the latest FSA studies, offering a dual perspective from both technical and applied standpoints. It also connects FSA with other disciplines such as information systems and finance, reviewing the principles of financial forecasting and supporting the market predictability of financial sentiment from a financial theory perspective. The survey discusses the relationship among market sentiment, investor sentiment, and financial textual sentiment, highlighting their interdependencies. It also reviews the scope of FSA studies, the challenges in FSA techniques and applications, and the interactive relationship between FSA techniques and applications. The survey identifies six areas that cause FSA failure, including irrealism mood, rhetoric, dependent opinion, unspecified aspects, unrecognized words, and external reference. It also discusses the evaluation metrics used in FSA, including regression and classification metrics, and the methods used in FSA, such as lexicon approaches, machine learning approaches, and deep learning approaches. The survey concludes that FSA is a critical area of research with significant implications for financial markets and decision-making.