Centuries of experience allow us to assume some fundamental properties about the universe. Firstly, the universe (=nature) is real. For science, the world that our senses identify is the only real one. We no longer debate whether it is ideals that are real and what we see and experience mere shadows of this reality, or whether this world is but illusion or some dreamer's cosmic dream. Plato's allegory of the cave, or the notion that the world in which we live is but "dream time" have been relegated (along with other myths and speculations) to the philosophical realm, and excluded from the scientific one.
Secondly, the universe is consistent. The universe behaves the same way, any time, any place, for any one. There are no special cases, no exceptions, no miracles, no special people to intervene with magic, wizardry or special knowledge. Science is not the domain of a special priesthood or mystical initiations. Our experiences today are the same as Newton's and Einstein's experiences. They are repeatable.
Thirdly, the universe is reasonable. This word has two meanings, and both apply here. The first is that the universe behaves in a logical, orderly fashion. The second, is that the human mind is capable of reasoning it out. The world is neither too mysterious, nor are we too stupid to understand it.
The domain of science is limited to the real, the natural, the sensory. Science deals with all the objects and phenomena that are accessible to our senses and to the various instruments that expand these senses. Because these phenomena are universal, they are accessible to anyone. And because the universe is consistent, they are accessible over time and space, they are repeatable. While speculation about astrology, dowsing, Nostradamus, people being abducted by aliens in UFOs, or the Bermuda Triangle might be fun, they are not within the domain of science.
How do scientists arrive at those systematic descriptions and explanations, and how do these become incorporated into the accepted mainstream of science? Usually, the sequence of steps associated with scientific methodology is listed as follows:
1. collecting data (=making observations); 2. formulating a hypothesis (often defined as an educated guess); 3. testing the hypothesis (doing experiments); 4. elevating it to a theory (less iffy than a hypothesis) 5. testing the theory (doing more experiments) which eventually, when proven, 6. the theory becomes a law; and 7. this law allows predictions which lead to new observations, etc.
This is a wonderful algorithm but, the most polite thing that can be said for this model is that any resemblance to how scientists really do science is purely coincidental. It just is NOT done that way.
They do NOT begin with observations as the above litany would suggest, Even the simplest object in the world, the simplest event, is so complex, so multifaceted, that is is impossible to know how, or what part to observe or describe unless it is within a context. Nor do scientists do science as a matter of course, like a mechanic would maintain an aircraft to keep it flying. They become engaged in the process of doing science when they perceive that there is a reason.