Modeling data sets contain many biases, but some are less obvious than others. In this course, you will learn how to identify and measure Subjectivity bias, which often times appears in various use cases that deal with both structured and unstructured data. This type of bias, also known as 'Observer' bias, appears in such applications as Sentiment analysis and Survey responses. In the first case, the sentiment of a given entity is prone to subjectivity of the person reading the content. In the second case, the person or persons filling out the survey might also use their subjective judgement without hard evidence to support their answer to a particular question. Once the subjective responses are made in both of the above situations, they are often used as inputs to predictive models, i.e. as either dependent or independent variables. The lack of objective accuracy in such modeling data is what becomes known as ‘Subjectivity’ or ‘Observer’ bias. What is the potential damage of such bias to the model outcome/predictions? Is there any way such bias can be measured beforehand to mitigate its impact? What are some of the assumptions and limitations that need to be considered when handling this type of bias in real-life applications? These are the questions that will be answered in this course from both practical and theoretical perspectives.