Informatics skills

Communicating, structuring, questioning, searching, and making decisions

Informatics is not just about systems and data. It is also about how we communicate, how we structure information, how we ask questions, how we search efficiently, and how we make decisions with uncertainty.

Common problems

  • Ambiguous messages
  • Missing context or assumptions
  • Information overload (too much detail)
  • Searching the wrong space, or searching inefficiently
  • Decisions driven by intuition, not evidence

A practical workflow

  1. Communicate with the audience in mind
  2. Structure content so it can be used
  3. Question to remove uncertainty
  4. Search strategically (keywords, space, algorithms)
  5. Decide using probabilities and preferences (utility)

Quiz

Drag or click an answer. Keyboard-friendly too.
Progress: 0/5 correct
❓ Which maxim is mainly about “being relevant”?
Drop or click the correct answer
❓ Which structure best fits a “how-to” guide?
Drop or click the correct answer
❓ Which question best reduces ambiguity?
Drop or click the correct answer
❓ Which search strategy guarantees the shortest path in an unweighted graph?
Drop or click the correct answer
❓ What does “utility” represent in decision theory?
Drop or click the correct answer
Communication fails when people have different assumptions, different mental models, or different priorities. Informatics improves communication by making meaning explicit.

Grice’s maxims

Quantity

Give the right amount of information (not too little, not too much)

Example: If someone asks “What changed?”, do not paste 3 pages of raw logs. Summarise and link to detail.

Quality

Be truthful and evidence-based

Example: “This model is accurate” becomes “This model achieved AUC 0.78 on held-out data.”

Relation

Be relevant to the goal

Example: In a safety incident, start with impact and actions taken, not the full history of the project.

Manner

Be clear, avoid ambiguity, present in an organised way

Example: “It broke” becomes “The import failed at row 312 due to an invalid date format.”

Message rewriter (quick practice)

Improved version
This is a scaffold: it prompts clarity (what happened, where, impact, and next step). You can adapt it to SBAR, tickets, or structured updates.
Structure is a tool for making information easier to find, understand, and use. Different structures fit different goals.

Knowledge-oriented

Explains concepts (what and why)

Task-oriented

Supports actions (how to)

Placeholder-oriented

Fills a template (standard fields)

Structure builder

Suggested structure
Tip: Structure is a “decision” about what to include, what to exclude, and what order improves use.
Good questions reduce uncertainty. They uncover assumptions, clarify goals, and improve the quality of models, requirements, and decisions.

Question types

Clarifying questions

Define terms, success criteria, and scope

Examples:
  • What does “better” mean here?
  • What is the decision this will support?
  • What is in scope vs out of scope?

Probing questions

Understand causes, constraints, and drivers

Examples:
  • What usually happens before this issue appears?
  • What constraints (time, staff, access) matter most?
  • What is the biggest source of uncertainty?

Testing questions

Challenge assumptions and check evidence

Examples:
  • What evidence supports that?
  • How would we know if this fails?
  • What alternative explanation fits the data?

Prioritising questions

Decide what matters most right now

Examples:
  • Which outcome matters most to users?
  • What is the minimum viable version?
  • If we can only fix one thing, what is it?

Turn a vague request into a good question

Better questions you can ask next
Strong questioning is how we “debug” understanding before we build models, systems, or interventions.
Searching is not only “typing into Google”. In informatics we often search a space of possibilities: states, paths, hypotheses, documents, or solutions.

Search strategy simulator

Search result
BFS explores level-by-level. DFS goes deep early. Heuristic search prioritises “promising” states.

Quick Boolean query builder

Suggested query
Tip: Use AND to narrow, OR to expand synonyms, and quotes for phrases.
Decisions should combine:
  • Beliefs (probabilities)
  • Evidence (data and inference)
  • Preferences (utility: what we value)

Bayes’ theorem calculator

Posterior = how likely something is after seeing new evidence.

Base rate (0 to 1)
True positive rate
False alarm rate
Posterior probability
Base rates matter. A “good test” can still produce many false positives when the prior is low.

Expected utility chooser

Expected utility combines probability and value: choose the option with the higher expected score.

Option A

Option B

Decision recommendation
If you change “utility”, you change what the “best” decision is; even if probabilities stay the same.

Communicating

  • Be clear, relevant, truthful, and the right length
  • Make assumptions explicit
  • Adapt to audience and goal

Structuring

  • Match structure to purpose
  • Use templates to standardise outputs
  • Prioritise findability and reuse

Questioning

  • Clarify definitions and scope
  • Probe constraints and uncertainty
  • Test assumptions with evidence

Searching & decision making

  • Pick the right search strategy for the space
  • Use probability + evidence to update beliefs
  • Use utility to represent what outcomes matter