D606 Data Science Capstone, catalog number DTAN 6218, is the three CU capstone that integrates the WGU Master of Science, Data Analytics core with the three Data Science specialization courses, requiring you to evaluate an organization's needs and opportunities and then act on that evaluation with the full toolkit. It is the course where the degree is assessed as one thing rather than as a sequence of separate courses.
Integration is the whole assignment
Capstone returns rarely happen because the analysis was weak. They happen because the submission is a good project that could have been produced after two courses instead of after the whole degree. Integration is not a bonus in D606. It is the thing being measured.
Practically, that means the project has to visibly draw on the preparation and statistics from the core, on the modelling from the specialization, and on the communication and deployment thinking that surrounds them. A capstone that ends at a validated model, with no account of how the result would reach a decision maker or survive contact with production, has skipped the integration the title promises.
The second scored dimension is organisational fit. The catalog frames this as evaluating needs and opportunities, and that framing is deliberate. You are expected to justify why this problem was worth an organisation's money, not merely why it was tractable with available data. An interesting dataset is not a business case.
Turning scored aspects into a section plan
Capstone scoring detail lives in your Course of Study, and capstone rubrics tend to carry more aspects than ordinary course rubrics. Count them before writing anything. Each is scored on its own against a three point scale and each needs a 2, so a project can be excellent in eight aspects and still be returned for the ninth.
Capstones also usually arrive in stages, with an approval step before the main work. Treat every stage as separately scored and do not carry forward an unapproved assumption. A scope that shifted after approval, without saying so, is a predictable source of trouble at the end.
The word budget, worked. Assume ten scored aspects and a report in the region of 4,000 words. Reserve 250 words for an executive opening and 200 for a close, leaving about 3,550, or roughly 355 words per aspect. Then rebalance deliberately: the organisational need aspect and the results interpretation aspect each go to 500, funded by trimming the method description aspects to about 280. A capstone that spends more words describing what it did than justifying why it mattered reads as a technical report rather than as a graduate capstone.
A structure that fits a data science capstone
Your capstone directions and any supplied template take precedence over everything below. Where the shape is yours to choose, this order supports the integration the course asks for.
| Section | What belongs in it | How it gets read |
|---|---|---|
| Executive summary | Problem, approach, headline result and recommendation, in one page | Written last, read first, and often the only part a stakeholder reads |
| Organisational need | Who has the problem, what it costs them, what changes if it is solved | The aspect that makes a capstone a business project rather than an exercise |
| Data and preparation | Sources, quality, governance limits, preparation with reasons | Governance and privacy limits are expected at graduate level |
| Methodology | The analytical approach and why it fits both the data and the decision | Should visibly draw on more than one prior course |
| Results | Findings with uncertainty attached, against a stated baseline | Certainty without uncertainty is the classic capstone overreach |
| Implementation | How this reaches production and who operates it after you leave | The integration aspect most often left thin |
| Evaluation plan | The metric that will show whether the solution worked, measured when and by whom | Distinguishes a delivered project from a finished assignment |
| Limitations and ethics | What the analysis cannot support, and who could be harmed by acting on it | Expected, brief, and conspicuous by absence |
Write the executive summary last and then check it against the body line by line. Capstones drift during a long build, and a summary written from memory of the original plan is a common mismatch.
Evidence craft at capstone scale
A capstone runs long enough that evidence discipline decides whether the final week is assembly or archaeology.
- Keep a decision log from day one, recording each analytical choice, its date and its reason. Most of the justification writing at the end comes straight from it.
- Version the data. A capstone that quietly changed its dataset in week six and never said so has an unreproducible result.
- Attach uncertainty to every headline figure. Confidence intervals or ranges are expected at graduate level.
- Record the provenance and permission status of any organisational data, and say plainly if it was synthetic or public.
- Cite external evidence for the business case, not just for the methods. Industry benchmarks make the need section defensible.
- Use APA throughout and keep the reference list building as you go rather than reconstructing it at the end.
The strongest capstone sentence is usually a limitation stated in the writer's own voice, followed by the reason the work still supports the recommendation. It reads as authority rather than as hedging.
What separates Competent from a submission sent back
Capstone aspects are scored independently, and with more aspects in play the chance that one is thin rises accordingly.
- The problem is an organisational problem with a cost attached, not a dataset that happened to be available.
- Results carry uncertainty and are compared to a baseline the organisation would recognise.
- Implementation and evaluation both exist as real sections rather than as closing paragraphs.
- The work visibly integrates several courses, and the report says where each contribution came from.
- The executive summary matches the body exactly, including the numbers.
Capstone performance assessment work can be revised and resubmitted with no grade penalty, so a return is a schedule event rather than a grade event. It is still the most expensive return in the degree, because capstone evaluation queues are long and terms are six months at a flat rate.
Six mistakes that cost time in D606
- Choosing the project for the data rather than for the need. It produces a capstone that cannot answer why this mattered.
- Scope drift after approval. If the project changed, say so and say why, rather than hoping the change goes unnoticed.
- Ending at the model. Implementation and evaluation are integration aspects and they are where the capstone earns its name.
- No uncertainty on headline numbers. A single point estimate presented as fact is the classic overreach.
- Reconstructing justifications at the end. Reasons recalled months later are vague, and vagueness is scoreable.
- Writing the summary first and never revisiting it. After a long build it usually describes a project you no longer did.
Pacing a capstone inside a six month term
A capstone fails on schedule far more often than it fails on quality, so plan it as a calendar problem first. Terms at WGU run six months at a flat rate, and a capstone that overruns does not just arrive late, it consumes a second term of tuition for one course.
Work backwards from the end. Reserve the final three weeks entirely for evaluation, response and any resubmission, because that window is not yours to control and the queue at term end is the busiest of the year. Reserve the two weeks before that for assembly: executive summary, reference list, appendices, formatting and the line by line check that the summary matches the body.
That leaves the build. Split it at the approval point, which is the first hard checkpoint in most capstones. Everything before approval is problem definition, data access and a defensible business case, and it should be finished early because nothing downstream can safely start until it is settled. Everything after approval is analysis, results, implementation and evaluation planning, and it is the part most likely to expand.
Two scheduling rules save more time than any writing technique. Do not begin the analysis until the data is in hand and inspected; a capstone that discovers in week eight that the promised dataset lacks the key variable has to restart. And write each section as it completes rather than saving writing for the end, because reasons recorded while fresh are specific and reasons reconstructed in month five are vague. Vagueness is exactly what an aspect scores down.
How support works on this course
Send the rubric from your Course of Study, the capstone directions and any template. What comes back is aspect mapped and staged: help shaping a defensible problem statement, a structure that carries integration through every section, results written with uncertainty attached, and implementation and evaluation sections that actually exist, plus a walkthrough so you can defend the project as your own.
The capstone is the course where pace planning matters most. Terms run six months at a flat rate, and a capstone started in month five is a capstone finished in a term you have to pay for twice.
Questions students ask about D606
Is D606 the same course as DTAN 6218?
Do I need an employer to sponsor my capstone project?
Can you write the capstone for me?
Where D606 sits in WGU's programs
The July 2026 catalog places this code in 1 current WGU program. Open a program page for the complete standard path and term positions. The live Degree Plan remains authoritative after transfer credit, substitutions, and mentor planning.
The assessments, one by one
The public catalog does not publish this course's PA/OA identity or task count. WGU Tutors publishes at most one PA manual per course and only from a WGU-controlled public rubric. Until that source exists, PA help begins from the student's real Course of Study and OA support remains preparation only.