<p>Automated personality analysis is a relevant task in many areas of computational sciences, like natural language processing, human-computer interaction, and digital pedagogy. Handwriting is widely investigated to analyse emotional states and personality because it conveys important information about the physical, mental, and emotional states of the writer. The steps in a typical system are data collection and digitization, pre-processing, feature extraction, and finally classification. Classification methods vary from traditional rule-based ones to the latest deep learning techniques. The objective of the current work is to present a systematic survey on automated personality analysis from handwriting. Depending on the classification method, related works are divided into three categories, namely, rule-based classification, machine learning-based classification, and, deep learning-based classification. It is found that the rule-based classification techniques are being replaced by machine learning-based ones over time and the latest trend is shifting to deep learning. Moreover, graphology, the conventional tool for personality analysis from handwriting is being replaced by psychology-based models like the Minnesota Multiphasic Personality Inventory, Myers-Briggs Type Indicator, and the Five-factor model. The main difficulty in comparing the works under study is that most of them have used in-house datasets. Nonetheless, the best accuracy obtained for graphology with rule-based classification is 93.77%, and for graphology with machine learning is 98.00%. Besides that, the best accuracy for the personality model with traditional machine learning is 98.10% and the personality model with deep learning is 98.46%. The current challenges and future directions of research are discussed in detail.</p>