# Research paper starter kit

This kit holds everything you need to write an In Motion Automation research letter paper in LaTeX. You only need this folder and the links below. Nothing here depends on a private machine.

Download the whole kit: https://brand.agentsinmotion.ai/research/starter-kit/starter-kit.zip

## What is in the kit

| File | What it is |
|---|---|
| `TEMPLATE.tex` | The house preamble with an empty paper skeleton |
| `references.bib` | One placeholder BibTeX entry |
| `ima-research.mplstyle` | Matplotlib style with the cool-lane palette |
| `figures/make_example_fig.py` | Example Python figure script (EXAMPLE DATA) |
| `figures/make_example_fig.m` | Example MATLAB or Octave figure script (EXAMPLE DATA) |
| `figures/example_fig.pdf`, `.png` | Output of the Python script |
| `figures/example_fig_matlab.pdf`, `.png` | Output of the MATLAB script |

All numbers in the example figures are made up. Replace them before you ship anything.

## 1. Copy the template

1. Copy the whole folder and rename `TEMPLATE.tex` to your paper name.
2. Edit only the identity block near the top of the body: `\shorttitle`, `\attribution`, `\title`, `\author`, `\date`.
3. Keep the preamble as it is. Font, margins, header, footer, section look and the abstract style are locked.
4. Write the paper in the section skeleton. Add sections if the work needs them.
5. Put real sources in `references.bib` and cite them with `\citep{key}` or `\citet{key}`.

## 2. Build figures as vector PDFs

1. Copy `figures/make_example_fig.py` and swap in your data.
2. Load the style at the top of every script: `plt.style.use("ima-research.mplstyle")`.
3. Save a PDF for the paper and a PNG for decks and previews, from the same script.
4. Run it: `python3 figures/make_example_fig.py`
5. Include the PDF in LaTeX: `\includegraphics[width=0.85\textwidth]{figures/example_fig.pdf}`

MATLAB users can start from `figures/make_example_fig.m`. It also runs in GNU Octave: `octave --no-gui figures/make_example_fig.m`

## 3. Compile with latexmk

```
latexmk -pdf TEMPLATE.tex
```

latexmk runs pdflatex and BibTeX as many times as needed. Clean up with `latexmk -c`. You need a TeX Live install with the recommended fonts and the latex-extra collection (for tgtermes, titlesec, fancyhdr, natbib).

## 4. Dataviz rules

- Keep it sparse. No 3D, no rainbow, no shadows or glow.
- Use the cool-lane palette: ink `#10151C`, navy `#0B1622`, sky `#C5D4E2`, cyan `#4EC3E0`, muted `#5A6572`, grid line `#CDD5DD`.
- Orbit colors (blue `#3B6FE8`, violet `#6B4DE8`, magenta `#E04AA8`) are for one small highlight at most. Never fill a whole chart with them.
- Use real source data only, and name the source in the caption.
- Labels must not overlap. Give every axis a title with units.
- Use a real plotting library (matplotlib, MATLAB, Octave). No AI-drawn charts.
- No pie charts with more than four slices. No second y-axis unless the data needs it.
- Decks reuse the same figure files as the paper. Do not redraw them with new numbers.

## 5. Review and ship check

Before you export the PDF:

1. Run the ship check line by line: https://brand.agentsinmotion.ai/research/SHIP-CHECK.md
2. Read the prose against the house writing rules: https://brand.agentsinmotion.ai/research/HOUSE-WRITING.md
   Then check the paper against the paper standard, starting with its stranger test: https://brand.agentsinmotion.ai/research/RESEARCH-PAPERS.md
3. Send the paper to Chief (or the named reviewer) for PASS or HOLD. Build a deck only after the paper gets a PASS.
4. Keep the abstract to the question, the method and the findings. Keep budget lines and sourcing notes out of the abstract and the body.

## Links

- Research skill: https://brand.agentsinmotion.ai/skills/ima-research-documents/SKILL.md
- House writing: https://brand.agentsinmotion.ai/research/HOUSE-WRITING.md
- Paper standard: https://brand.agentsinmotion.ai/research/RESEARCH-PAPERS.md
- Ship check: https://brand.agentsinmotion.ai/research/SHIP-CHECK.md
- Brand guide: https://brand.agentsinmotion.ai/GUIDE.md
- This kit: https://brand.agentsinmotion.ai/research/starter-kit/