Paid PractitionerCertificate
Finding Relationships in Performance Data
What moves with what, and how often that question is answered falsely: correlation and its rank, distance and information-theoretic alternatives, the confounders and mediators that make the same arithmetic mean opposite things, and the decomposition that compresses a many-channel telemetry stream to the few modes carrying its energy.
- Lessons
- 8
- Quiz questions
- 64
- Exercises
- 1 (10 min)
- Calculators
- 1
- Time
- About 84 min
- Price
- €179+ VAT
- Certificate
- Graded final exam, then a PDF certificate with a number anyone can verify
- Who it's for
- Analysts who build or review performance models, choose input variables, or have to say whether an observed relationship in fleet data is real: anyone whose next step after cleaning a dataset is a correlation matrix.
- Assumed knowledge
- Our courses Describing Vessel Data and Cleaning the Signal, or equivalent. Correlation coefficients assume things about distributions that the first course teaches you to check, and they are meaningless on data the second course has not yet cleaned.
- Taught by
-
CEO & Founder, Blue Autonomy
- Naval architect and marine engineer, National Technical University of Athens
- 15 years of experience in fleet performance and operational data
- Chief Scientific Officer, NEMO, previously DeepSea Technologies, StormGeo, Athenian Sea Carriers, PwC
Part of the Performance Data Science specialisation.
What you will be able to do
By the end of the course. Each outcome names the lessons that deliver it, and the final exam samples every one of them.
- State the three properties of a correlation and what a coefficient does not contain, and recognise the two mechanisms that generate strong correlations between unrelated quantities: common cause and coincidence. Lesson 1
- Compute and test a Pearson correlation, separating significance from strength, and explain why large samples make almost any coefficient significant and almost none useful. Lesson 2
- Choose between Pearson, Spearman and distance correlation for the dependence at hand, and read the gaps between the three coefficients as a diagnostic of relationship shape. Lesson 3
- Interpret mutual information as uncertainty removed, and use it for feature screening while respecting its estimation difficulties. Lesson 4
- Compute a partial correlation and distinguish the confounder case from the mediator case on physics rather than arithmetic; recognise the proxy variables performance data leans on and document the conditions under which a substitution stays valid. Lessons 5, 6
- Run a principal component analysis on standardised multi-channel telemetry, read the loadings back as physics and choose how many components to keep; apply rank-k truncation of a snapshot matrix as a reduced-order model, and size a sensor array against spatial aliasing. Lessons 7, 8
Here is a result from a real vessel, one of two sister VLCCs of about 317,000 DWT whose performance recorders log one reading a minute. Across 380,443 cleaned records, speed through water and shaft power correlate at 0.868, which is the strong relationship physics leads you to expect between two quantities tied by a cube law. Control for propeller revolutions and fuel consumption, and the same correlation becomes −0.135: not merely smaller, but the other way round.
Two readings are available. One is that the relationship was never there. The other is that it is so thoroughly there that the controlled variables have already carried all of it, and removing them removed it. The arithmetic cannot tell you which, and choosing wrongly is how a variable that matters gets dropped from a model. Why the remainder should point the other way is a further question again, and lesson five takes it apart.
This course is about that class of problem. What is inside: the three properties of a correlation and the significance test that turns a coefficient into a claim; Spearman's rank correlation and the distance correlation for relationships that are not straight lines; mutual information, which needs no assumption about shape at all; partial correlation, and the difference between a confounder and a mediator that decides what partialling out means; proxy variables, which work until the day they do not; and principal component analysis and the singular value decomposition, which answer the question for twenty channels at once and hand back a handful of modes that carry the physics.
The opening lesson is free to read as a sample. By the end you should be able to take a cleaned dataset, produce a defensible account of what depends on what, and explain to a sceptical reader why the relationships you kept are real and the ones you discarded were not.
What the course covers
- What Correlation Is, and What It Is Not 8 min
- Pearson, and What Its Significance Test Actually Says 10 min
- When Pearson Will Not Do: Spearman and Distance Correlation 9 min
- Mutual Information: Dependence Without a Shape 8 min
- Partial Correlation: The Confounder and the Mediator 23 min + exercise
- Proxies: Variables That Work Until They Do Not 8 min
- Many Variables at Once: PCA and the Singular Value Decomposition 9 min
- Modes, Energy and Reduced-Order Models 9 min
How was this course?
You have completed every lesson. One rating and, if you like, one line: it goes to the author, and it is how the course gets revised.
€179 + VAT
Paid once, for one person, for as long as the Academy runs: every lesson, the Excel workbook, the final examination and the certificate. Part of the Performance Data Science specialisation, which costs less than its courses bought one at a time.
Buying online needs JavaScript switched on in your browser. Otherwise write to info@blueautonomy.gr and we arrange access.
You own this course. Every lesson below is open to you.
Common questions
What exactly is included?
8 lessons (about 84 minutes in total), 64 quiz questions with explained answers, and 1 working Excel calculator: Correlation and partial correlation.
How much does it cost?
€179 plus VAT, paid once: €221.96 with 24% VAT, which is what an individual in the EU pays; outside the EU no VAT is added, and a company in another EU country with a VAT number valid in VIES pays the net price. That buys every lesson, the Excel workbook, the final examination and the certificate, for one person, for as long as the Academy runs. The course is also part of the Performance Data Science specialisation, which costs less than its courses bought one at a time. Everything else is on the pricing page.
Is there work to do, or is it only reading?
1 lesson carries a required exercise, about 10 minutes of hands-on work in total. You download that lesson's Excel workbook, follow the numbered steps against the numbers the lesson states, then change one input and watch the answer move. The exercises are where the method stops being something you have read and becomes something you can run; the stated course time includes them.
Do I get a certificate?
Yes. The course ends with a graded final exam, marked on our server rather than in your browser: you see your score and which questions you got wrong, but not the answers, because the questions are drawn from a pool. Pass it and you get a PDF certificate in your name, with a certificate number on it.
That number is the point. Anyone you show the certificate to can check it themselves at academy.blueautonomy.gr/verify, with no account and no email to us: the page states the name, the course and the date we issued it. A certificate nobody can verify is a decorated PDF, and we would rather issue one that stands up.
Can I preview the course?
Yes: the first lesson, What Correlation Is, and What It Is Not, is free to read in full. It is a real lesson, not a trailer: same depth, same quiz.
Do I need prior knowledge?
Our courses Describing Vessel Data and Cleaning the Signal, or equivalent. Correlation coefficients assume things about distributions that the first course teaches you to check, and they are meaningless on data the second course has not yet cleaned.