Cognitive biases can significantly influence decision-making processes, even when using data, as they are inherent errors in thinking shaped by cultural and personal experiences. Data, although perceived as objective, can still reflect human biases, which are further complicated when machine learning models inherit these biases, leading to unexpected and potentially harmful outcomes. The article discusses six common types of data bias: confirmation bias, selection bias, historical bias, survivorship bias, availability bias, and outlier bias, each affecting how data is interpreted and decisions are made. To mitigate these biases, strategies such as recording pre-analysis beliefs, ensuring representative sampling, acknowledging historical biases, considering both successful and unsuccessful outcomes, focusing on broader trends, and investigating outliers are recommended. Recognizing and addressing these biases is crucial for making more informed and accurate decisions in data-driven environments.