A common advice is to first think about what you personally would like to create. While you could certainly go the route of going with what is popular and what-not, you should at least show some interest/appreciation for the topic, being it the genre, the theme, etc.
So for example, let us say you would like to create a visual novel to focus on telling a story (maybe with some gameplay), or alternatively a game focused on 3d environments and fast gameplay. Based on that (and more of course) you can judge further what the best approach would be.
Personal questions: How much time/effort can/do you want to spend on it, what could you achieve and by what means? While it takes passion and talent to write a good novel, it requires less focus and talent in environmental design and complex gameplay mechanics. Wanting to create a game is not equal to wanting to >create< the game. And sometimes it just will not work out and it is better to accept this going forward and settle with something else you would want to make. (A different genre or playstyle in this case).
Technical questions: Are you already capable of coding, if not would you be willing to learn? Some people swear on creating their own engine for their games, but it is certainly not for everyone.
Depending on the answers above, you can now think about what game you want to set up and by what means (the platform, engine, alone or in a team, etc.)
In regard to engines: Unreal is not always the best choice. It is a very powerful and versatile engine, but it can easily be too much (in many ways) and overshoot the goal. Usually there are alternatives depending on what you would want to do. (For example there are dedicated engines focusing on 2D games, platformers, rpgs, visual novels, etc.). That being said, Unreal can be a choice, it is just a matter of knowing what your game will be and where it should end up.
The next step would be learning the ropes of the engine (or how to code depending on where you want to go). There are usual tutorials - either in text form or videos - and it is best to try what you learn as you go. Another common advice here is to not start with your dream game (or at least not in its 'final form') but rather play around with what you can do and improve over time. Releasing updates (or finished projects) to the puplic can motivate you and be a learning experience. Some will also mention game jams in that regard, as they can be seen as a motivational boost to produce something and learn by doing. It will also help you getting into the public in terms of people knowing your work.
And from there, you 'just' have to improve. Get better at the implementation of your work technically, improve designing your game (in every aspect) and gather people who are interested in what you do.