The workshop aims to introduce participants to methods of data analysis using AI tools in an environment that provides control over data, code, work history, and project access. Particular attention will be given to working with Codex in a local environment, using a local repository and GitHub to manage versions, document the work, and monitor the analytical process. Participants will gain practical knowledge of how to prepare a secure working environment, organise data, conduct analyses, and document each stage so that the results can be reproduced and their quality can be checked.

During the course, participants will learn how to use Codex to support the analysis of numerical, textual, visual, and multimodal data. They will learn how to design a repository structure, separate input data from code and results, protect sensitive data, use configuration files, write instructions for an analytical agent, and conduct an iterative data analysis process. The workshop will cover data preparation, exploration, cleaning, analysis, visualisation, and reporting.

The course presents both the benefits and limitations of AI tools. It also includes demonstrations and practical exercises using participants’ own data or sample datasets. Participants will practise creating a secure workflow in which analyses are performed locally, code and documentation are version-controlled, and protected data are not transferred to external services without proper control. Particular attention will be given to working with confidential data, anonymisation, pseudonymisation, access control, change history, and the careful use of GitHub in research and organisational projects.

The course is intended for researchers, analysts, doctoral students, and anyone who works with data and wishes to use AI in a structured, secure, and controlled way. It will be particularly useful for people who analyse data from surveys, documents, interviews, scans, images, tables, archival materials, digital resources, or multimodal datasets while also needing to ensure data security, follow institutional procedures, and maintain a reproducible record of the analytical process.

Upon completion of the workshop, participants will receive a certificate of completion by email.

Date: XX XX 2026, 9:00 a.m.–2:00 p.m. (online workshop).

Time: 5 hours, including a coffee break.

Location: online.

Group: up to 15 persons.

Teacher: Joanna Siwek, Ph.D.

Price: 200 EUR (plus VAT if applicable) per person.

Methods of payment: wire transfer.

Topics:

1. Introduction to Codex
2. Data security principles
• Sensitive, confidential, and personal data
• Anonymisation and pseudonymisation
• Local work and cloud-based work
• Data access control
• Files excluded from the repository
• Secure management of keys, passwords, and configuration data
3. Preparing the working environment
• Running Codex locally
• A local project repository
• Folder structure
• Instruction files for Codex
• Documenting analytical assumptions
• Git and GitHub in data analysis
4. Managing an analytical project
• Code versioning and working with branches
• Quality control of results
• Reproducible analysis
• Documenting analytical decisions
5. Data preparation
• Importing data from different formats
• Data cleaning
• Data validation
• Describing the data structure
• Creating data dictionaries
• Separating raw data from processed data
6. Numerical data analysis
• Exploratory data analysis
• Descriptive statistics
• Identifying missing data, outliers, and errors
• Grouping and comparing data
• Visualising numerical data
• Automating repeated analyses
7. Text data analysis
• Preparing a text corpus
• Cleaning and normalising text
• Categorising and coding content
• Identifying patterns and topics
• Summarising large collections of texts
• Analysing documents and source materials
8. Visual data analysis
• Working with scans, images, and visual materials
• Describing image features
• Extracting information from visual materials
• Combining visual descriptions with metadata
• Checking the quality of image interpretation
9. Multimodal data analysis
• Combining numerical, textual, and visual data
• Designing a shared data structure
• Comparing information from different sources
• Creating reports that combine different types of data
• Checking the consistency of results
10. Reporting results
• Generating analytical reports
• Creating charts and tables
11. Analytical agents
• Writing custom instructions for Codex
• Creating an agent to support data analysis
• Defining tasks and limitations
• Checkpoints and error alerts
• Working iteratively with code, data, and documentation

Learning Outcomes

Upon completion of the course, participants will be able to:

1. Explain the potential of Codex and other AI tools in the analysis of numerical, textual, visual, and multimodal data.
2. Prepare a local environment for secure data analysis.
3. Design the structure of a local analytical repository.
4. Use Git and GitHub to manage versions of code, documentation, and analytical results.
5. Identify which data may be stored in a repository and which should be excluded.
6. Apply basic data security principles, including anonymisation, pseudonymisation, and access control.
7. Use Codex to explore, clean, analyse, and visualise numerical data.
8. Use Codex to work with large collections of texts, documents, and source materials.
9. Analyse visual data such as scans, images, maps, diagrams, and documents.
10. Combine different types of data in multimodal analyses.
11. Create repeatable and reproducible analytical workflows.
12. Document assumptions, decisions, code, and analytical results.
13. Create instructions and agents that support specific analytical tasks.
14. Check the quality of results produced with the support of AI.
15. Prepare reports, charts, tables, and summaries based on analytical results.
16. Use AI tools to improve the efficiency of analytical work while maintaining data security.

Joanna Siwek, Ph.D.

Dr Joanna Siwek is a researcher in the Department of Artificial Intelligence at the Faculty of Mathematics and Computer Science of Adam Mickiewicz University in Poznań. Her research focuses on the development of artificial intelligence algorithms, particularly mobile robot swarms equipped with empathic reasoning capabilities. She also investigates applications of fuzzy sets in finance and decision theory.

Dr Siwek is the head of the Computer Science degree programme at the Faculty. Her teaching activities encompass cloud computing, applications of fuzzy sets, the integration of business and economics with computer science, and artificial intelligence in the context of artificial empathy. She collaborates with Poznań University of Technology, Poznań University of Economics and Business, and the Institute of Computer Science and Cognitive Science in Catania, Italy.