## Stop Searching, Start Researching How much time did you lose last month searching for a specific file? A script from that one analysis, the "final" version of a figure, or the raw dataset a collaborator sent you six months ago? If you’re like most researchers, the answer is probably "too much." The chaos of poorly organized project folders is more than just a minor annoyance; it’s a direct threat to your productivity and the reproducibility of your work. Many graduate students and early-career researchers develop their data organization habits reactively, leading to a messy tangle of folders and confusingly named files like `Analysis_Final_FINAL_v2.xlsx`. This ad-hoc approach doesn't scale. As projects become more complex and collaborative, a lack of structure can lead to costly errors, wasted time, and the dreaded inability to reproduce your own results months later. This guide provides a simple, robust framework for organizing your digital research workspace. By implementing a consistent folder structure and a clear file naming convention, you can create a system that makes your work easier to find, understand, and share. It’s a small investment of time upfront that pays massive dividends in clarity and efficiency for years to come. ## The Foundation: A Sensible Folder Structure The core principle of good data organization is to separate project components by their function. A well-designed folder structure makes it intuitive to find what you need and prevents the accidental modification of critical files. While the exact structure can be tailored to your field, a great starting point looks like this: ``` Project_Name/ ├── 01_data/ │ ├── raw/ │ └── processed/ ├── 02_code/ │ ├── functions/ │ └── exploration/ ├── 03_analysis/ │ ├── figures/ │ └── tables/ ├── 04_manuscripts/ │ ├── draft_01/ │ └── submission/ ├── 05_admin/ │ ├── proposals/ │ └── meetings/ └── README.md ``` Let's break down why this works. ### Why a Numbered Prefix? The numeric prefixes (`01_`, `02_`, etc.) aren't just for show. They force your operating system to keep the folders in a logical, chronological order, reflecting the typical research workflow from data to publication. Without them, your folders would sort alphabetically, placing `admin` before `data`, which makes less intuitive sense. ### The Five Core Folders 1. **`01_data`**: This is the home for all your data. The crucial step here is to create two subfolders: * **`raw/`**: This folder should be treated as a vault. Place your original, untouched data files here. Once a file is in this folder, it should never be modified. This is your ground truth, ensuring you can always return to the starting point. * **`processed/`**: Any cleaned, transformed, normalized, or otherwise manipulated data goes here. If you merge two raw datasets, the result goes here. If you correct errors in a file, you save a new version here. This clean separation is the cornerstone of a reproducible workflow. For more on this, check out our guide on [creating a reproducible research package](/blog/how-to-create-a-reproducible-research-package-that-journals-love/). 2. **`02_code`** (or `02_scripts`): All scripts used for data processing and analysis live here. This includes R scripts, Python notebooks, Stata do-files, or any other code. Keeping code separate from data and results makes your project much easier to navigate and version control. If you're new to version control, our [beginner's guide to Git for researchers](/blog/git-your-research-together-a-beginners-guide-to-version-control/) is a great place to start. 3. **`03_analysis`**: This folder holds the outputs of your code. When your scripts in `02_code` run on the data in `01_data`, they should save their results here. * **`figures/`**: All plots, charts, and visualizations. * **`tables/`**: Formatted tables, summary statistics, and model outputs. 4. **`04_manuscripts`**: This is where you do your writing. It can contain Word documents, LaTeX files, or even markdown files managed in the [Alfred Scholar manuscript editor](https://alfredscholar.com). You can create subfolders for different drafts, journal submissions, or conference presentations. 5. **`05_admin`**: A catch-all for the project's administrative files. This includes grant proposals, IRB documentation, meeting notes, and budget spreadsheets. It’s the project management hub that keeps the non-research parts organized. ### The Most Important File: `README.md` At the top level of your project, you should always have a `README.md` file. This is a simple text file that acts as a guide to the entire project. A good README should briefly explain: * The goal of the project. * The contents of each main folder. * Any specific instructions needed to run the code. * The source of the raw data. * Contact information for the project lead. Creating this file at the start of a project and updating it as you go will save you—and anyone you collaborate with—countless hours. It’s the internal documentation that makes your project understandable a year from now. ## A Naming Convention That Works for You (and Your Computer) A good folder structure is only half the battle. If your files are named unpredictably, you'll still be wasting time. The goal of a file naming convention is to make files both human-readable and machine-sortable. ### Rules for Better File Names 1. **No Spaces or Special Characters**: Avoid spaces and special characters like `!@#$%^&*?`. They can cause problems with command-line tools and scripts. Instead, use underscores (`_`), hyphens (`-`), or CamelCase. 2. **Use ISO 8601 Dates (YYYYMMDD)**: Always put dates at the beginning of a filename to ensure chronological sorting. `20260804_LabMeetingNotes.md` will always appear before `20260805_LabMeetingNotes.md`. 3. **Be Descriptive but Concise**: The name should give you a good idea of the file's content without being a full sentence. Keep filenames under about 30-35 characters. 4. **Include Version Numbers**: For files that undergo revision, add a version number at the end. Use leading zeros (e.g., `_v01`, `_v02`) so that `v10` sorts correctly after `v09`. ### Putting It All Together: A Practical Example Let's imagine you're working on a project called "MitoCell" studying mitochondrial function in cell lines. **Bad file names:** * `data from august 4.csv` * `Final plot for paper.png` * `My Code (revised).R` * `MitoCell_Paper_final_comments_from_jane_FINAL.docx` **Good file names:** * `20260804_MC_CellLineA-RawData_v01.csv` (in `01_data/raw/`) * `20260809_MC_Metabolism-Boxplot_v03.png` (in `03_analysis/figures/`) * `20260807_MC_Process-Raw-Data_v02.R` (in `02_code/`) * `20260815_MC_Manuscript-Draft_v04.docx` (in `04_manuscripts/`) This system provides a wealth of information at a glance: the date, the project, a brief description, and the version number. It's instantly searchable and sortable. ## Don't Forget Backups: The 3-2-1 Rule A perfect organization system is useless if your hard drive fails and you lose everything. Data backup is not optional; it’s a fundamental part of research data management. The most widely accepted best practice is the **3-2-1 Backup Rule**. * **3 Copies**: Keep at least three copies of your data. This means your primary working copy and two backups. * **2 Media Types**: Store the copies on at least two different types of media (e.g., your computer's internal hard drive and an external hard drive). This protects you if one type of storage technology fails. * **1 Off-site Copy**: Keep at least one copy in a physically separate location. This could be a cloud service like Dropbox or Google Drive, or your institution's server. This is your safeguard against a localized disaster like fire, flood, or theft. Automating this process with cloud sync and scheduled backups to an external drive is the best way to ensure it actually gets done. Losing data is a catastrophic, and entirely preventable, setback. By adopting these simple practices—a logical folder structure, a consistent file naming convention, and a robust backup strategy—you can build a research workflow that is efficient, reproducible, and far less stressful. You'll spend less time wrestling with files and more time making discoveries.