SciSpace Academy
Tutorial

How to Find, Analyse and Communicate Open Research Data with AI

Dr. Lyndon WalkerEnglishData analysis

TL;DR

1-min read

Open and public datasets offer researchers a quick route to quantitative evidence, but finding the right data, analysing it correctly and reporting it clearly all take time. This tutorial follows that full journey in SciSpace, from locating data to producing publication-ready summaries, and closes with the research integrity questions that come with using AI on data. It is designed for quantitative researchers and graduate students who want a repeatable, transparent process for secondary data analysis.

Key takeaways

  • Open research data are datasets published by researchers, institutions or governments for reuse, often under licences that allow analysis and redistribution, such as national surveys, repository deposits linked to articles and public health statistics.
  • SciSpace can help you discover and retrieve relevant open datasets as the first step of the workflow.
  • Multi-step statistical analyses can be run within a single AI-assisted workflow, which keeps the process in one place.
  • Interpreting the output carefully and identifying meaningful insights is what turns analysis into conclusions, and raw statistical output can then be converted into structured summaries for publication.
  • Reusing open data requires respecting licence terms and citing the original source.
  • Before analysing data with AI, check institutional and national data policies, especially for sensitive or restricted data, keep a record of each step, verify outputs independently and disclose AI assistance in your methods.
Apply it

To apply this, choose an open dataset relevant to your question, run your analysis guided in SciSpace, and document each step so the work aligns with your data governance requirements.

How to do it

  1. Define the research question and the variables, population, timeframe, and granularity required.

    Define the question and required variables, population, timeframe, geography, unit, and granularity before searching. This prevents choosing a convenient dataset that cannot answer the question.

  2. Use SciSpace to find papers and data-availability statements that point to relevant repositories.

    Use papers and data-availability statements to identify repositories, then verify the repository owner and persistent identifier. Record alternative datasets considered.

  3. Review the metadata, codebook, license, citation instructions, and data-quality documentation.

    Review the license, codebook, collection methods, provenance, citation instructions, missingness, and known quality limitations before download or upload.

  4. Download or upload the permitted data, create a cleaning log, and run an approved analysis plan.

    Create a versioned cleaning log and an analysis plan that matches the dataset’s design and permitted use. Preserve original variables and document derived fields.

  5. Export code, tables, and charts and write a limitations-aware summary with the dataset citation.

    Export code, tables, and charts with the dataset citation, version, and license. State representativeness, missing data, measurement limits, and any restrictions on interpretation.

Watch the tutorial38 min

How to Find, Analyse and Communicate Open Research Data with AIPlay video: How to Find, Analyse and Communicate Open Research Data with AI

Chapters

  1. Welcome and speaker introduction
  2. Can you put research data into ChatGPT? Data governance and ethics
  3. Rules for using AI in data analysis: verify, prompt well, use domain knowledge
  4. How AI slop can damage your academic career
  5. What can SciSpace do? Tour of the agent homepage and AI tools
  6. How to do an AI literature review with SciSpace Deep Review
  7. What is forward and backward citation searching in AI literature review?
  8. Chat with PDF: ask AI to explain text, maths and graphs in a paper
  9. How to export references from SciSpace to BibTeX or RIS
  10. How can I use AI for data analysis? Upload a CSV to SciSpace Agent
  11. Complete a half-finished dataset with AI web scraping, then analyse it
  12. Why you must still verify AI statistical analysis
  13. How to run a CAPM beta analysis with AI using Python yfinance
  14. Can AI find a dataset for you? Analysing WHO open data with SciSpace
  15. How to write a detailed AI prompt for statistical analysis
  16. Is SciSpace data analysis transparent? Code, data, plots and report
  17. How to use AI for data analysis without uploading confidential data
  18. Can PhD students use AI for research? What AI should do vs what you do

Try this on SciSpace

Frequently asked questions

What is open research data?

Open research data are datasets that researchers, institutions or governments publish so others can reuse them, usually under a licence that permits analysis and redistribution. Examples include national surveys, repository deposits linked to journal articles, and public health statistics. Reusing them enables secondary analysis and replication, provided you respect the licence terms and cite the original source.

Is it acceptable to analyse research data with AI tools?

Often yes, but check your institution's and country's data policies first, especially for sensitive or restricted data. Keep a record of each analysis step, verify statistical outputs independently, and disclose AI assistance in your methods. The tutorial discusses these governance considerations alongside the practical workflow so integrity is built in from the start.

Related tutorials