- You can have about 4 figures in an average paper for a journal. Figure out how to tell the story of the experiment in pictures alone. You cannot possibly spend too much time making the figures pretty & logically consistent, look at examples from the journal (or class) you are submitting the paper to for proper style & formatting.
- Describe the pictures/graphs, how you got them, and what they tell you. This is 75% of the paper.
- Write the setup & procedure you followed to do the experiment, put this before the descriptions. Write down the analysis you followed afterwards. This is usually straightforward for a class project, you probably have a lab manual.
- Harder, write an introduction/motivation for the experiment (including the background theory) - why did you do it, and what did you expect to find?
- Write a conclusion - what did the data tell you, and how does that contrast with what you expected? Can be short.
- Write an abstract/lead paragraph. This is the hard part, since most people will read your abstract first and *then* decide whether to read your paper.
- Hardest, come up with a catchy title. No one will read your abstract, much less your paper, unless you catch their eye first.
- Make sure you properly acknowledged your funding agents and included as co-authors everyone who performed significant work. Being highly political, leave this to a senior project member if possible ;-)
Anyway, that's what I tell my grad students, and what I follow myself by and large, and it seems to work well enough. Sort of an inverted pyramid - start with the trivial things you can describe easily, and work your way up to the crucial aspects everyone will notice first once you've figured out everything else. Scientific writing isn't much different than any other writing. You have to know your audience, capture the reader's interest, and tell your story in an efficient and compelling fashion.