Data & Analytics

Intentionally malformed frontmatter. Not a real skill.

Explains what a member or a role can do in a PostHog project, using the access control MCP tools. Use when the user asks what someone can see or edit, who can edit dashboards or feature flags, why a member can or cannot open a dashboard, notebook or table, what a role grants, which properties are hidden from someone, or how the project's default access is set. Covers what each level means, how the stored rule, the enforced level and the inherited access relate, what the null values mean, whic...

Extract Claude Code session analytics to CSV, Markdown, or HTML files

Claude Code session analytics - post stats to PRs/issues, extract to CSV

Ensure gradient zeroing happens before backward pass, not after. Verify forward pass returns tensors requiring gradients.

Force proportional analysis when emotional responses fail to scale with problem magnitude

Scoring framework using Reach, Impact, Confidence, and Effort to objectively prioritize product features and projects

Forecast by comparing to similar past cases (outside view) instead of analyzing unique project details (inside view) - cures planning fallacy

Online learning updates models continuously as new data arrives, one sample at a time or in mini-batches, without retraining from scratch

Systematic methodology for choosing the most appropriate machine learning algorithm based on problem characteristics, data properties, and constraints

The tendency to perceive past events as more predictable than they actually were, leading to distorted learning and overconfidence

Problem-solving strategy that makes the locally optimal choice at each step with the hope of finding a global optimum

Visually map all potential causes of a problem across structured categories to systematically identify root causes through team collaboration

Systematic techniques for transforming raw data into features that better represent the underlying problem structure, improving ML model performance

Systematic problem-solving approach that breaks down complex problems into smaller, independent subproblems, solves them recursively, and combines their solutions

Build consensus expert forecasts through iterative anonymous surveys with controlled feedback to minimize groupthink and anchor biases

Product becomes smarter and more valuable as it collects more usage data from users, leveraging machine learning to create indirect value

Express uncertainty as a plausible range instead of false precision - quantify how confident you are in estimates to avoid overconfidence in decisions

Identifies the minimal computational structure required to predict a system's behavior from limited observations

Align predicted probabilities with actual outcomes - train yourself so things you say are 70% likely actually happen 70% of the time

When improving forecast accuracy requires balancing how well predictions match reality vs. how well they distinguish between outcomes

Measure forecasting accuracy by calculating mean squared error between predicted probabilities and outcomes when improving calibration through systematic feedback

Machine learning inference pattern where predictions are generated for large datasets at scheduled intervals rather than in real-time

Ignoring statistical base rates in favor of vivid case-specific information when assessing probability