1 of 7

Open Science in data collection and analysis

06/05/2025

SciLifeLab Training Hub

Ineke Luijten, PhD

2 of 7

How to implement Open Science

The Open Science workflow

Plan &

Design

Collect &

Analyse

Publish &

Spread

Access &

Reuse

Evaluate &

Build

Communicate & Involve

Reuse data from data repositories

Use open-source software for method documentation & data analysis

Publish preprints

Publish Open Access

Publish all research outputs

Practice Open peer review

Deposit all outputs in open repositories

Use open licensing

Attach persistent identifiers

Add rich metadata

Use responsible research metrics

Adopt qualitative research assessment

Track Open Science contributions

Share key insights through (social) media

Encourage Citizen Science

Turn research into MOOCs or OERs

Preregister hypotheses, study protocols, analysis

workflows

Create a data management plan

3 of 7

How to implement Open Science

The Open Science workflow

Plan &

Design

Collect &

Analyse

Publish &

Spread

Access &

Reuse

Evaluate &

Build

Communicate & Involve

Reuse data from data repositories

Use open-source software for method documentation & data analysis

Publish preprints

Publish Open Access

Publish all research outputs

Practice Open peer review

Deposit all outputs in open repositories

Use open licensing

Attach persistent identifiers

Add rich metadata

Use responsible research metrics

Adopt qualitative research assessment

Track Open Science contributions

Share key insights through (social) media

Encourage Citizen Science

Turn research into MOOCs or OERs

Preregister hypotheses, study protocols, analysis

workflows

Create a data management plan

4 of 7

How to implement Open Science

Collect &

Analyse

The Open Science workflow

Collect &

Analyse

Document methods and analyze data with open-source tools

Work according to Good Research Practice standards

Reuse data from data repositories

5 of 7

Collect &

Analyse

Method documentation

What: provide a clear, detailed, and accessible record of how your data was collected, processed, and analyzed

Why: others can understand, evaluate, and reproduce your research decisions

How: Protocols.io (sharing and collaboratively editing research protocols)

Jupyter Notebooks (mixes code, results and narrative)

Github (version-control)

Open Science Framework (bombine protocol, data, and code in one public space)

eLabFTW (open-source ELN, inventory and experiment tracking)

Openness in methodology

Image Protocol by NutJannah Noun Project CC-BY 3.0

Speaker notes:

Now it’s time to put in to practice the tools that you have chosen to work with in the plan & design phase of your project, and that you have described in your data management plan.

For documenting methods, use open tools to document them in a transparent and reusable way. Transparent documentation means providing a clear, detailed, and accessible record of how your data was collected, processed, and analyzed. The goal is to ensure that others (and your future self) can understand, evaluate, and reproduce your research decisions—not just the results. So: document your analysis pipelines!

Documentation is your future-proofing tool. Tools like Jupyter Notebooks, RMarkdown, or Protocols.io let you document not only what you did, but how and why you did it. This makes your process understandable, reproducible, and usable by others—or even by yourself, months later.

Share data in real-time using electronic lab notebooks (ELNs) or Open platforms like eLabFTW, Jupyter Notebooks, OpenLabArchives, OSF or GitHub.

Background:

Why Is It Important?

Supports reproducibility – If methods aren't fully documented, your results can't be verified or reused.
Increases trust – Transparency makes it easier to spot errors, biases, or assumptions.
Enables reuse – Clear methods make your research more useful for meta-analyses or follow-up studies.
Facilitates collaboration – Good documentation reduces onboarding time for new collaborators or team members.

Purpose: tool examples

Code + narrative: Jupyter, RMarkdown, Quarto

Versioning: GitHub, GitLab, OSF

Protocol sharing: Protocols.io, OpenWetWare

Metadata standards: DataCite schema, Dublin Core, ISA-tab

Notebooks & lab logs: eLabFTW, OpenLabArchives, LabArchives

6 of 7

Openness in data analysis

Collect &

Analyse

Analyse data with open-source tools

What: use tools that are freely available and transparent in their functioning to process, visualize and analyse data

Why: no expensive licenses required, exact replication of analyses, fast-evolving

How: OpenRefine (clean and transform messy survey or archival datasets )

R/Python (scripting language for data wrangling, statistics, and visualization)

OpenSesame (tool for building behavioral experiments)

JASP (graphical user interface but still open and reproducible)

Select interoperable formats

What: use non-proprietary file formats that are platform-independent

Why: data can be read and reused across platforms, tools, and time

How: CSV (Comma-Separated-Values for tabular data)

JSON (JavaScript Object Notation for hierarchical or structured data)

XML (eXtensible Markup Language for structured, machine-readable data with metadata)

TXT, Markdown, YAML (for plain-text documentation and config files)

Image file format by Fauzan AdiimaNoun Project CC-BY 3.0

Speaker notes:

Whether it’s R, Python, JASP, or OpenRefine, using open-source tools keeps your workflow transparent and accessible. Proprietary software can lock you—and your collaborators—out of full control or reproducibility. Plus, open tools often have vibrant communities and better long-term support.

Choosing open, interoperable formats like CSV, JSON, or XML ensures our data can be read and reused across platforms, tools, and time. Proprietary formats might work now, but what about 10 years from now? Open, interoperable formats are platform-independent and can be opened with a wide range of software tools, both now and in the future.

Background:

Open source tools:

What?�Use open-source tools—like R, Python, JASP, OpenRefine, or OpenSesame—to process, visualize, and analyse your data.

These tools are freely available, transparent in their functioning, and often support scripting, which improves reproducibility.
They cover a wide range of use cases—from statistical modelling and data cleaning to qualitative analysis and experiment design.

Why?

Transparency: Anyone can inspect how a tool works, which avoids black-box results.
Reproducibility: Scripts and code ensure that analyses can be exactly replicated.
Collaboration: Open tools make it easier to share work without requiring expensive licenses.
Longevity & Community: Popular open tools are maintained by large user communities and often evolve faster to meet new research needs.

Non-proprietary file format:

What?

CSV (Comma-Separated Values) – for tabular data
JSON (JavaScript Object Notation) – for hierarchical or structured data
XML (eXtensible Markup Language) – for structured, machine-readable data with metadata
TXT, Markdown, YAML – for plain-text documentation and config files

These formats are platform-independent and can be opened with a wide range of software tools, both now and in the future.

Why?

Longevity – Proprietary formats (like .sav from SPSS or .xlsx from Excel) may become unreadable or unsupported. Open formats can be used indefinitely.
Accessibility – Your data can be opened and reused across different systems and software without needing expensive or specific tools.
Machine-readability – Interoperable formats are easily parsed by programming languages and tools (e.g., Python, R, or APIs).
Data sharing & reuse – Repositories, collaborators, and future users can understand and integrate your data more easily if it’s in a common, open format.

How?

Export your data from proprietary tools into CSV or JSON before sharing or archiving.
Avoid features that aren’t preserved across formats (e.g., pivot tables, conditional formatting).
Validate your files by opening them in different tools (e.g., Excel, R, Notepad) to check consistency.
Use metadata standards (e.g., README files, JSON sidecar files) to document your file structure, variable definitions, and units.
Automate format conversion with scripts (e.g., Python’s pandas.to_csv() or json.dump() functions).

7 of 7

Open Science in practice

Reuse data from repositories�Saves time and resources

Find well-documented datasets in trusted repositories, check licensing and ethics, and clearly cite and document how you use them.

Good Research Practice�Ensures ethical, credible outcomes

Conduct research that is rigorous, transparent, ethical, and fully documented so that it can be understood, verified, and reused by others.

Collect &

Analyse

Speaker notes:

Everything we’ve discussed so far—choosing open formats, documenting transparently, using open tools—isn’t just about following trends. It’s about upholding the standards of Good Research Practice. When you design your data collection and analysis to be open, you’re also making your research reproducible, auditable, and more valuable to the scientific community. This opens the door to reuse—by yourself, your collaborators, or researchers you’ve never met.

GRP means being:

Rigorous – using validated, transparent procedures.
Responsible – documenting choices, reducing bias, avoiding unnecessary duplication.
Respectful – making efficient use of time, resources, and data—your own and others’.

In this session, we will learn more about how working in line with GRP principles in data collection and analysis automatically increases compliance with Open Science principles.

Furthermore, we tend to focus on making our own data open and FAIR—but that’s only half the equation. Reuse is equally essential. Open science encourages us not to start from scratch every time, but to build on what’s already been done. Reusing existing data allows us to ask new questions, combine sources for meta-analyses, or validate earlier findings. It’s a way of being more efficient, more collaborative, and ultimately more responsible with our research resources. Of course, reusing data requires critical thinking—checking documentation, assessing data quality, understanding consent and licensing—but these are all part of the modern research skill set. Today’s session will give you tools and strategies for doing this effectively.

Background:

What is GRP?

Good Research Practice (GRP) refers to a set of principles that guide responsible, ethical, and high-quality research conduct. It ensures that research is:

Rigorous – well-designed, methodologically sound, and evidence-based.
Transparent – fully documented, openly shared where possible.
Reproducible – able to be verified and replicated by others.
Ethical – conducted with integrity and in compliance with legal, ethical, and disciplinary standards.

Why does GRP matter?

Builds trust – With funders, peers, the public, and future collaborators.
Prevents errors and misconduct – Clear documentation, transparent workflows, and good data management reduce the risk of mistakes and questionable practices.
Improves reproducibility – Others can verify your work or build on it, advancing knowledge more efficiently.
Supports open science – GRP aligns directly with the FAIR principles (Findable, Accessible, Interoperable, Reusable).
Protects researchers – Ensures proper attribution, tracks contributions, and supports defensible decision-making.

How do you work according to GRP?

Plan your research transparently�Use pre-registration platforms or protocols.io to document your study design, hypotheses, and data collection methods in advance.
Use open, interoperable formats�Store and share data in formats like CSV or JSON to ensure longevity and compatibility.
Track and document everything�Use version control (e.g., GitHub or OSF) to track changes in data, code, and analysis decisions.
Apply open and validated tools�Analyze your data using tools like R, Python, OpenRefine, and JASP that allow full reproducibility and auditability.
Store your data securely and ethically�Protect sensitive data and use proper consent and licensing to enable ethical reuse.
Share your work responsibly�Deposit your data and code in trusted repositories (e.g., Zenodo, OSF, Dataverse) with rich metadata and clear documentation.

Reusing data�What is it?

Reusing data means incorporating existing datasets—collected and shared by other researchers—into your own research. These datasets are often published in open repositories like Zenodo, OSF, Figshare, UK Data Service, or Dryad.

It can involve:

Re-analyzing data to answer new questions
Combining datasets in meta-analyses or comparisons
Validating or replicating published findings

Why reuse data?

Efficiency – Saves time and resources; no need to collect new data if it already exists.
Ethics – Reduces participant burden and research duplication.
Impact – Builds on existing knowledge and increases the visibility of shared data.
Transparency – Strengthens scientific integrity by enabling replication and cross-checking.
Training – Allows students and early-career researchers to practice analysis on real-world data.

How to reuse data effectively

Find datasets in trusted repositories�Use platforms like Zenodo, OSF, Figshare, UK Data Service, or institutional archives.
Check for licensing and consent�Ensure the data is legally and ethically reusable (e.g., Creative Commons licenses, de-identified data).
Assess data quality and metadata�Review documentation to understand the dataset’s context, structure, and limitations.
Cite the dataset properly�Always acknowledge the original creators using the DOI or recommended citation format.
Document your reuse transparently�Record how you accessed, processed, and interpreted the data to maintain reproducibility.