The A&V SIG exists to grow practical, deep understanding of analytics and visualization at the Netherlands eScience Center, both in terms of concepts and real implementations.
We do not aim for black-box usage. We aim for the level where you can explain what is happening, spot when a method is a bad fit, and implement or adapt solutions with confidence. The sweet spot is where mathematics, statistics, and good visual design reduce complexity and make results easier to trust, interpret, and communicate.
We meet roughly every 4 weeks. Occasionally we run shared sessions with other SIGs when topics overlap.
If you want to present, co-host a session, or invite an external speaker, please add yourself to the schedule via a pull request or open an issue in the repo.
- Jesse Gonzalez
- Peter Kok
| Date (YYYY-MM-DD) | Type (A, V, A&V) | Speaker | Topic | Notes/Links |
|---|---|---|---|---|
| YYYY-MM-DD | A&V | TBA | TBA | TBA |
| YYYY-MM-DD | A | TBA | TBA | TBA |
| YYYY-MM-DD | V | TBA | TBA | TBA |
| YYYY-MM-DD | A&V | TBA | TBA | TBA |
- Statistical modelling and inference
- Uncertainty quantification and error analysis
- Time series, forecasting, and anomaly detection
- Bayesian workflows
- Causal inference
- Dimensionality reduction and clustering diagnostics
- Experiment design and evaluation
- Visual encodings, perception, and common pitfalls
- Data storytelling for scientific results
- Uncertainty visualization
- Interactive dashboards and exploratory tools
- Visual analytics (analytics + interactive visuals)
- Geospatial visualization
- Network visualization
- Scientific visualization (e.g., fields, volumes, meshes, point clouds)
- Turning messy outputs into something humans can understand
- Making uncertainty visible instead of hiding it
- Debugging models with visuals
- Building analysis tools that are usable by non-experts
Typical formats (we mix them):
- Talk + discussion (internal or external speaker)
- Hands-on session (small exercises, live coding)
- Paper or video club (short prep, deeper discussion)
- Show and tell (demos of tools, plots, dashboards)
- Clinic session (bring your problem, we dissect it)
If you are stuck on an analytics or visualization problem, bring it in. We can use part of a meeting as a mini-consulting session.
Built something cool, useful, or surprisingly effective? Bring it.
Examples:
- A visualization that revealed a model bug
- A plot redesign that changed how results were interpreted
- A pipeline that made analytics reproducible and explainable
- A dashboard that people actually use (rare species, please exhibit)
Open an issue or contact the SIG organizers to schedule a slot.
Use this list as inspiration. Add more via PR or issue.
- Model checking and diagnostics that people actually do
- Bayesian thinking without the religious wars
- Causal inference: when it works, when it does not
- Practical uncertainty quantification
- Time series: forecasting vs storytelling
- Uncertainty visualization patterns and anti-patterns
- Color scales that do not lie (and do not break for color blindness)
- Visualizing high-dimensional data without pretending it is 2D magic
- Networks: how to avoid hairball charts
- Geospatial: projections, distortions, and misleading maps
- Visual analytics: interactive tools for model exploration
- Communicating uncertainty to non-specialists
- Using visualization to debug data pipelines
- From notebook plots to production-grade dashboards
Be kind, be curious, and ask naive questions early. The goal is shared understanding, not showing off.
If something is unclear, it is probably unclear for 5 other people too. Ask it.