I recently started an internship in a private research company and I'm supposed to write an article for a conference in the next 6 months.
This is a process that confuses me. How do people look at some data and then come up with some ideas that they find worth publishing?
My work involves applying complex science measurements on some environmental data. There are around 40 measurements of complexity that I found so far, with 4 types of entropies. How is one supposed to believe that they have a proper understanding of all of them, have a proper understanding of the existing research and also come up with some novel insights?
Some content about the scientific process in general or about personal experiences would be highly appreciated.
Usually you have a general topic you like to talk about, but you can't because it is all well known. So you have to give it just a little twist.
This can be a new method you try, for example a neural network that does some thing. You then explain the general topic, then give some Infos about the NN and may compare its success to an established method. This has been done in every field now, that people already get sick of it. But this generally holds, just apply an out of field method and compare. It makes for interesting content and you can regurgitate the known theory to get enough mass.
If you have hard data, that is not easily available, because the process to receive it is espensive, you can publish with just some basic evaluation. Just make it a tiny bit more shiny.
In your case with environmental modelling you should be in between the two. Just keep it simple and well struktured.
The scientific process is the art of adding a lot of very small steps. You contribution can be tiny if you already walk a well known path.
It's all luck and knowing people in the end anyway.
https://en.wikipedia.org/wiki/Replication_crisis
When you add up all the ways that a scientific paper could be "wrong" it's clear that the median scientific paper is wrong.
In middle school they tell you a lot of things about the "Scientific Method" which are rarely practiced by real scientists. For instance there is the idea that you compare a "control" with an "experiment" condition but in the average scientific paper there is no "control". (e.g. even though they control background when they build a neutrino observatory, for instance, they can't run the detector in a neutrino-free environment)