- CDS here means Certified Data Scientist - Associate/Practitioner, issued by the Institute for the Certification of Computing Professionals (ICCP).
- The published Data Science Exam outline lists 12 official topics, but it gives no percentage weights.
- ICCP's store lists $350 per professional exam including proctoring and $249 per foundational exam; an older FAQ says $299.
- Active certification runs on a 3-year professional-development cycle: Associate 60 hours, Practitioner 120 hours.
Which CDS This Sheet Covers
"CDS" is a crowded acronym, and several unrelated credentials share it. This cheat sheet is about exactly one: the Certified Data Scientist - Associate/Practitioner, offered by the Institute for the Certification of Computing Professionals (ICCP). If you want the naming background first, see What Is CDS Certification? and What Does CDS Stand For?.
A cheat sheet is only useful if it is honest about its edges. Where ICCP's public material is clear, this page states the fact. Where it is not, this page says so rather than guessing. That matters because candidates routinely lose money and time by acting on a fee, date, or prerequisite that belonged to a different program.
The One-Page Fact Sheet
Here is the condensed reference. Anything marked "confirm with ICCP" is deliberately left open.
| Item | What is verified |
|---|---|
| Credential | Certified Data Scientist - Associate/Practitioner (CDS) |
| Issuer | Institute for the Certification of Computing Professionals (ICCP) |
| Exam outline | 12 official topics; no percentage weights or revision year published in the accessible outline |
| Professional exam fee | $350 per exam including proctoring (current store); an older FAQ lists $299 |
| Foundational exam fee | $249 per exam (current store) |
| Candidate application fee | $45 (generic ICCP fee) |
| Application materials | Resume and ethics agreement |
| Proctoring | ICCP exams use approved proctoring, including arranged remote or in-person options |
| Maintenance cycle | 3 years; Associate 60 hours and $35 annual fee, Practitioner 120 hours and $75 annual fee |
| Exam count, SKUs, prerequisites | Confirm with ICCP |
| Open-book, calculator, adaptive delivery | Not verified; confirm with ICCP |
All 12 Topics at a Glance
The outline organizes the field into 12 topics. They follow the natural arc of an analytics project: frame the question, get the data, prepare it, analyze it, model it, and communicate it reproducibly. Since no weights are published, treat every topic as examinable and do not skip any based on a rumored emphasis.
- Business & Technology Issues: Starting with the Question First (what problem are you trying to solve?)
- Data Storage, Big Data and Sources
- Mathematics and Statistical Data Science
- Programming Skills
- The Data Analytic Question and Types of Data and Reporting
- Tidying the Data: Data Cleaning and Quality
- Exploratory Analysis
- Statistical Modeling and Inference
- Prediction and Machine Learning
- Causality
- Written Analysis
- Reproducibility
For a deeper walk through each area, use the complete guide to all 12 CDS content areas.
Must-Know Concepts by Cluster
Rather than memorizing twelve disconnected lists, group the topics into four clusters. This mirrors how the questions interrelate: a prompt about modeling often assumes you already handled cleaning and framing correctly.
Cluster A: Framing and Data Foundations (Topics 1, 2, 5)
Business & Technology Issues and the Analytic Question
The outline's own subtitle, "Starting with the Question First," tells you the philosophy. Expect scenario-style thinking where the right answer is the one that clarifies the problem before reaching for a tool.
- Distinguish a descriptive, exploratory, inferential, predictive, and causal question; each demands a different method and a different claim.
- Know how the type of data (structured versus unstructured, categorical versus numeric, cross-sectional versus time-ordered) constrains what you can report.
- Be able to say what a stakeholder actually needs from a report, not just what is easy to compute.
Data Storage, Big Data and Sources
This topic is about where data lives and how it arrives.
- Relational versus non-relational storage and when each fits.
- What makes a dataset "big": volume, velocity, variety, and the practical consequences for processing.
- Evaluating a source for provenance, bias, and fitness for the question at hand.
Cluster B: Quantitative and Technical Core (Topics 3, 4)
Mathematics and Statistical Data Science
This is the conceptual backbone for everything later.
- Probability basics, distributions, and the logic of sampling.
- Summary measures of center and spread, and when each misleads.
- Linear algebra and calculus intuition as it supports modeling, without needing to derive everything by hand.
Programming Skills
Expect questions about reasoning with code, not trivia about one library.
- Data structures, control flow, functions, and reading short snippets to predict output.
- Manipulating tabular data: filtering, grouping, joining, reshaping.
- Writing code that others can run and understand, which links directly to the reproducibility topic.
Cluster C: Wrangling and Analysis (Topics 6, 7, 8)
Tidying the Data: Cleaning and Quality
- Missing values: recognizing the mechanism (random or systematic) before choosing deletion or imputation.
- Duplicates, inconsistent encodings, outliers, and unit mismatches.
- The principle of tidy data: one variable per column, one observation per row.
Exploratory Analysis
- Choosing visualizations that match the variable types and the question.
- Using exploration to generate hypotheses, not to confirm them. Know why that distinction protects you from false discoveries.
- Reading distributions, relationships, and anomalies from plots and summaries.
Statistical Modeling and Inference
- Estimation, confidence intervals, hypothesis tests, and what a p-value does and does not mean.
- Regression assumptions and how violations show up in diagnostics.
- Quantifying uncertainty and communicating it honestly.
Cluster D: Prediction, Causation, and Communication (Topics 9 to 12)
Prediction and Machine Learning
- Training versus test data, cross-validation, and overfitting versus underfitting.
- The bias-variance tradeoff and how model complexity moves you along it.
- Choosing evaluation metrics that fit the problem, especially with imbalanced classes.
Causality
- Correlation is not causation, and the specific reasons why: confounding, reverse causation, selection effects.
- Why randomized experiments support causal claims and what observational designs must do to approximate them.
- Recognizing when a predictive model is being misread as a causal one.
Written Analysis and Reproducibility
- Structuring a written analysis: question, data, methods, results, limitations, conclusion.
- Matching the strength of your language to the strength of your evidence.
- Version control, documented environments, fixed random seeds, and scripted pipelines so another analyst can rerun your work.
Key Takeaway
The topics build on each other. A strong answer on modeling usually depends on having handled framing and cleaning correctly, so study the clusters as a chain rather than as twelve islands.
Fees and Registration Mechanics
The fee facts that can be stated cleanly are these. ICCP's current store lists $350 per professional exam, including proctoring, and $249 per foundational exam. An older FAQ page shows $299 for the professional exam, so if you see two numbers online, you are likely seeing the older page and the current store. Treat the store as the live source and verify at checkout.
Separately, ICCP lists a generic $45 candidate application fee. The application itself calls for a resume and an ethics agreement. Exact CDS-specific requirements for education, experience, training hours, and references are not verified here, so confirm them directly with the issuer. The CDS requirements guide covers how to approach that conversation.
On delivery, ICCP exams use approved proctoring, including arranged remote or in-person options. The specific testing vendor and workflow for the CDS exam are not verified, which means you should ask ICCP what you need on exam day (identification, room setup, system checks) before you schedule. Timing and windows are discussed in the CDS exam dates article.
Staying Certified: The 3-Year Cycle
An active certification operates on a 3-year professional-development cycle. General ICCP tier rules specify:
| Tier | Professional-development hours (3-year cycle) | Annual fee |
|---|---|---|
| Associate | 60 hours | $35 |
| Practitioner | 120 hours | $75 |
Candidate-specific applicability should be confirmed with ICCP, since these are general tier rules rather than a personalized statement. The practical implication is that the credential is a commitment, not a one-time purchase. Start logging qualifying activity early rather than reconstructing it in year three. Whether that ongoing cost is justified is the subject of the CDS ROI analysis.
What Is Still Unverified
A trustworthy cheat sheet lists its gaps. As of this writing, the following are not confirmed from accessible public sources:
- Domain weights: the outline lists 12 topics but no percentages and no revision year, so any "Domain X is 20% of the exam" claim you encounter is unsupported.
- Which exam SKUs and how many exams the Associate or Practitioner tier requires.
- Open-book status, calculator rules, and whether delivery is adaptive.
- Passing score and pass rate. See the passing score article and pass rate discussion for what can and cannot be said.
- Exact education, experience, training-hour, and reference prerequisites.
The accessible Data Science Exam outline is also not a verified complete multi-exam pathway blueprint, so avoid assuming it maps one-to-one onto a single test. When you contact ICCP, bring this list as your question sheet.
Sequencing Your Review
Because there are no published weights, a sensible order follows dependency rather than rumor: build the foundation, then the analysis skills, then the judgment skills. This is a sample ordering, not an official schedule.
Frame and Source
- Topics 1, 2, and 5: question types, storage concepts, data types.
- Practice classifying sample scenarios by analytic question type.
Math and Code
- Topics 3 and 4: probability, distributions, then hands-on data manipulation.
- Do not skip this; later topics assume fluency here.
Clean, Explore, Infer
- Topics 6, 7, and 8: cleaning decisions, visual exploration, hypothesis tests and regression.
Predict, Explain, Communicate
- Topics 9 to 12: validation and overfitting, causal reasoning, written analysis, reproducible workflows.
- Finish with mixed practice questions across all twelve topics.
For the full preparation strategy, see the CDS study guide, and when you are ready to test yourself, work through the CDS practice tests. Candidates weighing effort should also read how hard the CDS exam is.
Where the Credential Gets Used
The skills the 12 topics describe map onto data analyst, data scientist, analytics engineer, and applied statistician roles, across any employer that works with data at scale. Industry-wide, that includes technology firms, financial services, healthcare organizations, government agencies, and consultancies. No specific employer demand figures or hiring statistics are asserted here, because none are verified for this credential. For how these skills show up in actual postings, browse the CDS jobs overview, and for compensation context see the CDS salary guide, which keeps to what can be supported.
The ICCP's emphasis on an ethics agreement at application also signals something about the credential's positioning: it is framed as a professional designation with conduct expectations, not merely a skills quiz.
FAQ
The Institute for the Certification of Computing Professionals (ICCP). It is distinct from other credentials that also abbreviate to CDS, so confirm you are on the ICCP pages when checking fees or requirements.
The accessible Data Science Exam outline lists 12 official topics, from business framing through reproducibility. It does not publish percentage weights or a revision year, so prepare for all twelve.
ICCP's current store lists $350 per professional exam including proctoring and $249 per foundational exam; an older FAQ shows $299. A $45 candidate application fee is also listed. Because the required number of exams for each tier is unconfirmed, no total is stated.
Open-book status, calculator rules, and whether delivery is adaptive are not verified. Ask ICCP directly before exam day rather than relying on forum answers.
Active certification uses a 3-year professional-development cycle. General ICCP rules specify 60 hours and a $35 annual fee for Associate, and 120 hours and a $75 annual fee for Practitioner; confirm how this applies to you.