Advice For Graduate Students in Statistics I have started my PhD in statistics this year, and I am looking for your best-practices, advice and (meta-advises) regarding how to grow and become a good academic researcher in the fields of statistics/ML.
General thoughts and links are welcomed, but in order to start the ball rolling, here are a bunch of questions gathered from Michael Steele's great article "Advice For Graduate Students in Statistics" (if I am missing important questions, or if some of the questions are meaningless - please also comment on it):


*

*Papers vs Thesis - how much should one focus on publishing papers during his PhD work?  How many papers should one realistically aspire to write?

*In what journals should one strive to get published in? (relevant questions link1, link2)

*How many hours a day should one spend on research (developing/dealing with your research question), and on learning (reading new papers/ attending courses)

*Where does one go to find "hot topic", or even better - a "soon to be hot topic"? (link1, link2)

*Once a "hot topic is found" how should one balance learning the basics of many aspect of the problem, with focusing on one aspect?


Obviously these questions are VERY general, and there are many angles for thinking/answering them - I hope to read your perspective on how to think about these general issues.
Thanks in advance!
 A: I recently finished my PhD, and I have some thoughts in regard to writing papers. I have to mention that I worked in the Dutch system where we have 4 years to complete our PhD, and I got a normal salary, competetive with industry starting salaries.
The norm in the Dutch system is to write around 4 papers in those 4 years. How many paper one can write also depends on the structure of your contract. If you have to follow many courses, or be a teaching assistent to get money, or have just 3 years, the number of papers is lower.
I did not write a traditional thesis, I used my papers as is and wrote an introduction and summary. I like this "collection of papers" approach because you focus on writing papers, not on a phd thesis that is probably not going to be read alot (at least for most of us :)). For a further career in science publications are important.
In terms of which journals to publish, I would say those you like reading yourself. Reaching an audience that shares your interest can get you citations. In terms of impact factor I would go for the highest IF that is still relevant for your field. I would also defintely consider open access journals such as those from the EGU. In this anyone, also from africa, can read your work without making some publisher rich... 
