In my opinion there are two things that you want to optimise as a researcher: Impact and Fun. Impact justifies your work as a valuable contribution to society and progress in sciences. Without impact it is difficult to have a successful career. Fun is essential to maintain your motivation. Without fun you would lose interest in your work and the drive to do research (which is, after all, an unusual and competitive type of work). In addition, in my opinion, one of the main advantages of being a researcher is precisely the freedom and flexibility to have fun while trying to have an impact. Fun may take many forms: reading a lot outside of your specific field of research, developing good code, brainstorming ideas with colleagues, teaching, public engagement and outreach, etc. Note that many of those may be very valuable contributions to your career (and perhaps even become your career at one point).
Historically, many significant contributions in the sciences have come people who combine a passion with the drive to make a difference (both can be interpreted very broadly). However, remember that you can be passionate and not necessarily work all the time. Work-life balance is not only critical for your well-being, but also for your work productivity itself. Furthermore, we all think, work, and have fun in different ways. The schedule and work habits that work best for your colleagues may not work at all for you (look at some of your favourite famous scientists, or even your own colleagues, who I am sure you will find have unusual routines). Discovering and honing your work style without feeling too pressured by others is a life long challenge.
In my opinion, the main criteria entering the choice of a project are:
- how short/long?: weeks, months, years?
- how risky?: are the data or code or model available, will the effect be actually detectable, etc
- how incremental?: obvious incremental extension of existing work versus a breakthrough discovery of a new effect or interpretation or even object.
- how collaborative?: do I have all the skills to complete it?
In astronomy, because of the availability of data and nowadays the usefulness of methodological innovations, there is space for pretty much all types of projects, although proper breakthroughs are obviously the most difficult to pursue.
How enjoyable and impactful the project will be depends on many subjective and field-specific factors. One is how valued (and in particular how visible and interpretable they are on your CV) your contributions will be (code, journal papers, short conference proceedings, talks, etc). In addition to the effect on your career, the choice of project should also definitely depend on the type and duration of the position you have. Long term fellowships allow you to spend more time on high-risk high-reward projects.
For each project, it helps to have (predicted) answers to those questions, and to revise that frequently.