How to Structure Research Data Folders

By Dr. Zubair Khalid, DVM, MS, PhD · 2026-08-27

Quick Answer

• Adopt a hierarchical folder structure that separates raw data, processed data, analysis scripts, and documentation to make files locatable and auditable.
• Name folders and files with consistent, descriptive conventions that include dates and version numbers so anyone on the team can navigate without asking.
• A folder structure alone does not guarantee data integrity, you must pair it with written naming rules, backup practices, and periodic reviews.

Why Folder Structure Matters for Research Data

Research projects generate files across multiple stages: raw instrument outputs, cleaned datasets, analysis code, figures, manuscripts, and correspondence. When these files accumulate without a deliberate organization system, locating a specific version becomes time-consuming and error-prone. A well-designed folder structure reduces the cognitive load of remembering where things live and supports reproducibility, collaboration, and compliance with funding requirements.

The National Library of Medicine hosts research methods resources that emphasize the importance of systematic documentation in biomedical research. Similarly, the EQUATOR Network provides reporting guidelines that assume researchers can locate and verify the data behind their published claims. If your files are scattered, you cannot easily demonstrate that your reported results match your actual analysis.

Funding agencies increasingly expect organized data practices. The NIH Data Management and Sharing Policy outlines expectations for planning how data will be managed and shared. While the policy does not prescribe a specific folder layout, it requires researchers to describe their data management approach, which is difficult to articulate if your file organization is ad hoc.

Core Principles of Research Data Organization

Separate Raw Data from Processed Data

Raw data are the original, unmodified outputs from your instruments, surveys, or observations. Processed data are cleaned, transformed, or derived versions. Keeping these in distinct folders prevents accidental overwriting of original records and makes it clear which files serve as the source of truth.

Create a raw folder that is read-only in practice. If you must modify a raw file, save the modified version elsewhere and document the change. This separation also simplifies audits because reviewers can compare processed results against the untouched originals.

Use Descriptive and Consistent File Names

File names should convey content without requiring the file to be opened. Include the project identifier, sample or subject identifier, date in ISO format (YYYY-MM-DD), and a version number. Avoid special characters, spaces, and overly long names that cause compatibility issues across operating systems.

A file named 2025-06-01_projectX_sampleA_raw.csv is immediately understandable. A file named final_v2_reallyfinal.csv is not. Consistency matters more than perfection, write down your naming convention and apply it uniformly.

Document Your Structure

A folder structure is only useful if people know how to use it. Place a README.md or STRUCTURE.txt file at the top level of your project folder that explains the purpose of each subfolder, the naming conventions, and any rules about file modification. This documentation becomes part of your research record and helps new team members orient themselves quickly.

The Recommended Hierarchical Folder Structure

The following structure separates data by processing stage and function. Adapt it to your specific research area, but preserve the core separation between raw data, processed data, analysis code, and documentation.

project_name/
├── 00_admin/
│   ├── proposals/
│   ├── ethics_approvals/
│   ├── funding_documents/
│   └── correspondence/
├── 01_raw_data/
│   ├── instrument_outputs/
│   ├── surveys/
│   └── observational_notes/
├── 02_processed_data/
│   ├── cleaned_datasets/
│   ├── derived_variables/
│   └── analysis_ready/
├── 03_analysis/
│   ├── scripts/
│   ├── figures/
│   └── results/
├── 04_documentation/
│   ├── protocols/
│   ├── data_dictionaries/
│   ├── README.md
│   └── file_manifest.txt
├── 05_manuscripts/
│   ├── drafts/
│   ├── figures_final/
│   └── submission_files/
└── 06_archive/
    ├── old_versions/
    └── completed_analyses/

Admin Folder

The 00_admin folder holds non-data project files such as grant proposals, ethics approvals, funding documents, and correspondence. These files support the research but are not part of the scientific analysis. Keeping them separate from data folders prevents clutter and makes compliance reviews easier when you need to show approval documentation.

Raw Data Folder

The 01_raw_data folder contains original, unmodified data files. Organize by collection method or source. Instrument outputs, survey responses, and observational notes each get their own subfolder. Files here should never be edited directly. If you need to correct an error in a raw file, document the correction in a processing log and store the corrected version in the processed data folder.

Processed Data Folder

The 02_processed_data folder contains cleaned and transformed datasets. Subfolders distinguish between cleaned versions of raw data and derived variables created during analysis. The analysis_ready subfolder holds the final datasets that feed directly into your statistical scripts. This separation lets you trace any analysis result back to the specific processed file and then to the raw source.

Analysis Folder

The 03_analysis folder holds your scripts, figures, and results. Scripts include statistical analysis code, data cleaning scripts, and any custom functions. Figures are generated plots and visualizations. Results include model outputs, tables, and summary statistics. Keeping code separate from data reduces the risk of accidentally modifying data files when running analyses.

Documentation Folder

The 04_documentation folder is the backbone of reproducibility. Protocols describe how data were collected. Data dictionaries define every variable name, its meaning, and its allowed values. The README file explains the overall structure. A file manifest lists every file in the project with its purpose and modification history. This folder answers the question, "What exactly did you do?"

Manuscript Folder

The 05_manuscripts folder tracks the writing and publication process. Drafts, final figures, and submission files each have their own subfolder. Version control for manuscripts is especially important because multiple authors may contribute revisions. Use clear version numbers and dates in file names to avoid confusion about which draft is current.

Archive Folder

The 06_archive folder stores old versions and completed analyses that are no longer actively used. Archiving instead of deleting preserves the full history of your research. If a question arises about a previous analysis, you can retrieve the archived files without rummaging through active folders.

At a Glance

Implementing the Structure in Practice

Step 1: Create the Top-Level Folders

Start by creating the seven top-level folders described above. Use a consistent naming convention with leading numbers so folders sort in logical order. The 00 prefix for admin and 01 through 06 for the remaining folders ensures that the structure displays predictably in file browsers.

Step 2: Write the README File

Create a README file at the project root that explains the folder structure, naming conventions, and rules for file modification. Include the project title, principal investigator or lead researcher, date of creation, and a brief description of the research. Update this file whenever you change the structure or conventions.

Step 3: Establish Naming Conventions

Decide on a naming convention before you create many files. A practical format is YYYY-MM-DD_description_version. For example, 2025-06-01_soil_samples_v1.csv. Record this convention in the README so all team members follow the same pattern.

Step 4: Migrate Existing Files

If you already have files scattered across drives and folders, migrate them into the new structure. Start with the most recent project and work backward. Do not attempt to reorganize everything in one session, allocate time each week until the migration is complete. As you move files, update the file manifest to reflect their new locations.

Step 5: Review and Adjust

After one month of using the structure, review how well it works. Ask team members about friction points. Adjust folder names or add subfolders as needed. The structure should serve your workflow, not the reverse.

Records and Measurements for Folder Health

File Manifest

Maintain a file manifest that lists every file in the project, its location, its purpose, and its modification history. This manifest serves as an index to your research data. Update it whenever you add, move, or delete files. A spreadsheet or a plain text file both work, the key is consistency.

Version Log

Keep a version log for processed data files and analysis scripts. Each entry should record the date, the person making the change, what changed, and why. This log creates an audit trail that supports reproducibility and helps resolve disputes about which version produced which result.

Storage Size Tracking

Monitor the total size of your project folder and the size of each subfolder. Rapid growth in the raw data folder may indicate that you are not archiving completed work. Unexpected growth in the analysis folder may indicate that scripts are generating large temporary files that should be cleaned up.

Backup Verification

Regularly verify that backups are working by restoring a test file from backup storage. A backup that has never been tested is not reliable. Record the date of each backup verification and the result.

Common Failure Patterns

The Everything Folder

Some researchers keep all files in a single folder, relying on file names to distinguish content. This pattern fails when the folder grows beyond a few dozen files because sorting and searching become inefficient. It also increases the risk of accidentally overwriting files with similar names.

The Nested Chaos

The opposite failure is creating too many nested subfolders, sometimes with only one or two files at the deepest level. Excessive nesting makes navigation slower and increases the chance of saving files in the wrong location. If a subfolder has fewer than three files, consider whether it needs to exist.

The Version Mess

Files named final, final_v2, final_v2_revised, and FINAL_ver3 indicate a broken versioning system. This pattern makes it impossible to know which file is current and which is obsolete. A version log and consistent version numbers in file names prevent this problem.

The Silent Overwrite

When multiple people work on the same project, someone will eventually overwrite a colleague's changes. This happens most often with processed data files and analysis scripts. A clear modification rule, such as "never edit a file that someone else created without saving a copy first," reduces this risk.

The Orphaned Data

Data files that are not referenced in any script or manuscript become orphans. They consume storage space and create confusion about whether they are still needed. Regular reviews of the archive folder help identify files that can be moved to long-term storage or deleted.

Welfare and Safety Context for Research Data

Data Integrity as a Safety Issue

In research involving animal subjects, human participants, or clinical samples, data integrity is a safety issue. A mislabeled or overwritten data file can lead to incorrect conclusions that affect subsequent experiments or clinical decisions. The folder structure described here supports data integrity by making it difficult to accidentally modify raw records.

Confidentiality and Access Control

Some research data are confidential. Patient records, personal survey responses, and proprietary industry data require restricted access. Your folder structure should include access controls at the operating system level, beyond organizational conventions. Store confidential data in folders with restricted permissions and document who has access.

Long-Term Preservation

Research funders and journals increasingly expect data to be preserved for years after publication. The NIH Data Management and Sharing Policy describes expectations for data preservation and sharing. A well-organized folder structure makes preservation straightforward because you can identify which files constitute the final dataset and which are intermediate work products.

Limitations of Folder Structures

Folder Structure Does Not Equal Data Management

A tidy folder structure is necessary but not sufficient for good data management. You also need documented protocols, version control, backup procedures, and a plan for sharing data after publication. The folder structure supports these activities but does not replace them.

No Universal Standard

The structure described here is a practical template, not a universal standard. Different research fields, institutions, and funding agencies may have specific requirements. The EQUATOR Network provides reporting guidelines that may include data availability expectations. Check whether your field or funder has specific requirements before adopting a structure.

Individual vs. Team Use

A folder structure that works for a solo researcher may need adjustment for a team. Teams need clearer rules about file ownership, modification, and versioning. They may also benefit from shared storage with access controls instead of individual local folders.

Tool Compatibility

Some analysis tools expect specific folder layouts. For example, certain statistical software packages have conventions for where data files and output files should live. Your folder structure should accommodate these tool-specific requirements while maintaining the overall separation of raw data, processed data, and analysis.

Professional Escalation Criteria

When to Involve Your Supervisor or Data Manager

Escalate to a supervisor or institutional data manager when you encounter any of the following situations:

• You discover that raw data files have been modified or overwritten and you cannot determine who made the change.
• You find that confidential data are stored in folders with overly broad access permissions.
• You are asked to share data for publication but cannot locate the final analysis-ready dataset.
• You inherit a project with a disorganized folder structure and need guidance on how to migrate it without losing files.

When to Consult Institutional Policy

Consult your institution's research data policy when you are unsure about retention periods, access controls, or sharing requirements. The NIH Grants and Funding pages describe expectations for funded research, and your institution may have additional requirements. Do not guess about compliance, ask the office responsible for research compliance.

When to Seek Ethics Guidance

If your data involve human subjects or animals, and you are unsure whether your data management practices meet ethical requirements, consult your institutional review board or ethics committee. The Committee on Publication Ethics describes core practices for research integrity that include responsible data handling. Ethics guidance is specific to your institution and jurisdiction, so seek local advice.

Practical Workflow for Setting Up a New Project

Before Data Collection

Set up the folder structure before you collect any data. Create all seven top-level folders and the README file. Write your naming conventions and modification rules. This preparation takes less than an hour and prevents disorganization from the start.

During Data Collection

Save raw data files directly into the appropriate raw data subfolder. Do not let files accumulate on your desktop or in a downloads folder. Record collection dates and any deviations from the protocol in the documentation folder.

During Analysis

Work in the analysis folder. Save scripts and results there. When you create a cleaned dataset, save it in the processed data folder and log the cleaning steps. Do not save intermediate files in the raw data folder.

Before Publication

Review the entire folder structure. Verify that the file manifest is current. Confirm that the analysis-ready dataset is clearly identified. Check that all figures in the manuscript have corresponding scripts and data files. This review ensures that a reviewer or reader can reproduce your results.

After Publication

Move the final dataset, analysis scripts, and documentation to the archive folder. If your funder or journal requires data sharing, prepare the sharing package from the archived files. Keep the archive intact for the retention period required by your institution or funder.

Options and Tradeoffs in Folder Design

Numbered vs. Unnumbered Folders

Numbered prefixes force a logical sort order but add visual noise. Unnumbered folders sort alphabetically, which may not match your workflow. Choose numbered prefixes if you want the folder order to reflect the research workflow.

Single Project vs. Multi-Project Structures

The structure described here is for a single project. If you work on multiple projects, create a parent folder for each project and replicate this structure inside each one. Do not mix files from different projects in the same folder tree.

Local vs. Shared Storage

Local storage gives you full control but makes collaboration difficult. Shared storage enables teamwork but requires clear rules about file ownership and modification. Many teams use a hybrid approach, with active work on shared storage and archives on local or institutional storage.

Cloud vs. On-Premises Storage

Cloud storage offers convenience and automatic synchronization but raises questions about data security and jurisdiction. On-premises storage gives you more control but requires you to manage backups yourself. Your institution may have policies about which storage options are permitted for research data.

Observations and Measurements for Continuous Improvement

Track Time Spent Locating Files

For one week, record how much time you spend searching for files. If you spend more than a few minutes per day, your folder structure or naming conventions need adjustment. This measurement provides concrete evidence for redesign decisions.

Monitor Folder Growth

Review folder sizes monthly. The raw data folder should grow steadily during data collection and then stabilize. The processed data folder should grow during analysis. The archive folder should grow as projects complete. Unexpected growth patterns indicate workflow problems.

Survey Team Members

If you work in a team, ask each member to describe where they would save a specific type of file. If their answers differ, your structure is not intuitive enough. Update the README and discuss the structure in a team meeting.

Review the File Manifest

Check the file manifest quarterly for accuracy. Files that are not listed, or listed files that no longer exist, indicate that the manifest is not being maintained. Reinforce the rule that the manifest is updated whenever files change.

Frequently Asked Questions

How do I name files so they sort correctly?

Use the ISO date format YYYY-MM-DD at the beginning of the file name, followed by a descriptive label and a version number. For example, 2025-06-01_soil_samples_v1.csv. This format sorts chronologically and makes the version immediately visible.

Should I keep raw data files read-only?

Yes, in practice. Set folder permissions so that only the lead researcher can modify raw data files, and even then, require a documented reason. If you must correct a raw file, save the corrected version in the processed data folder and log the change in the version log.

How many subfolders should I create?

Create subfolders only when they serve a clear purpose. A good rule is to create a subfolder when a folder would otherwise contain more than about twenty files of different types. If a subfolder would contain fewer than three files, consider keeping those files in the parent folder.

What should go in the README file?

The README should explain the purpose of each top-level folder, the naming conventions, the modification rules, and the versioning system. It should also list the project team members and their roles. Update it whenever you change the structure or conventions.

How do I handle files that do not fit any folder?

Create a misc folder for files that do not fit the standard categories, but review it regularly. Files that remain in misc for more than a few months should either be moved to an appropriate folder or archived. A growing misc folder indicates that your structure needs adjustment.

What is the difference between processed data and analysis results?

Processed data are cleaned or transformed versions of raw data. Analysis results are the outputs of statistical or computational procedures applied to processed data. For example, a cleaned survey dataset is processed data, while a regression table is an analysis result.

How long should I keep archived files?

Retention periods depend on your institution, funder, and field. The NIH Data Management and Sharing Policy describes expectations for funded research. Check your institution's research data policy for specific retention requirements. When in doubt, keep files longer instead of deleting them.

Can I use this structure for non-research projects?

Yes, the same principles apply to any project that involves data files, documents, and iterative work. The separation of raw inputs, processed work products, and documentation is useful for grant administration, clinical practice management, and even personal projects that involve data analysis.

Using the Evidence

Related Articles

• Bioinformatics Portfolio: Projects That Demonstrate Real Research Judgment
• Data Visualization for Biological Research: Choosing Figures That Preserve Meaning
• Git for Research Projects: Version Control for Code, Data, and Analysis Notes
• Reference Managers for Researchers: Choosing a Workflow That Survives Collaboration
• Open Science Practices: A Guide for Researchers

References and Further Reading

• Research Methods Resources. National Library of Medicine.
• EQUATOR Network. EQUATOR Network.
• Core Practices. Committee on Publication Ethics.
• NIH Grants and Funding. National Institutes of Health.
• ORCID for Researchers. ORCID.
• Data Management and Sharing Policy. National Institutes of Health.
• NCBI Literature Resources. National Center for Biotechnology Information.
• PubMed. National Library of Medicine.

This article is educational and does not replace institutional policy, professional advice, or applicable safety and regulatory requirements.