I would like to start studying ML & DS, but I feel I am a bit lost, so I don't really know what to study, what the prerequisites are, I mean I know I should study linear algebra, calculus, and statistics, but do not know the exact amount I need. I would be thankful if someone could provide me with a suitable roadmap to start, and an exact description of what I should do (What book to start with, whether to watch videos or read books .. etc)
The field(s) of ML and DS are very broad. Some topics are related to "classical" statistics (which require good knowledge in standard algebra and matrix calculation), while other fields expand into computer science and programming. Things like image classification (or "computer vision") for instance have been developed mostly in the computer science community, while other predictive models (linear regression etc) are widely used in "classical" statistics. As the fields/topics in ML/DS are very diverse, you will eventually need to focus on some fields according to your needs and interests.
In case you want to dive deeper into some topics, you may refer to "Elements of Statistical Learning", which is more formal and allows you to go into more details.
For more applied work related to neural nets, you may also follow the Keras Blog. Often links to various papers and other sources are provided, so that you can go into the details by topic if you are interested.
Start learning about the datasets, Data Cleaning, EDA (Exploratory Data Analysis), Visualization, some basics of linear algebra and statistics. Also learn libraries like pandas, numpy. These are some of the pre requisites before starting ML and DL.
For ML start with Regression and classification. Once you get grasp on them. Apply the learning in kaggle Competitions. You can also learn courses for free in kaggle. You will eventually learn the next steps as you move forward.
However, Hands-on Machine Learning with scikit learn, Keras and tensorflow by O'Reilly would be a great book for beginners as it covers almost all the topics in a practical approach.