🗂

Project requirements: the deliverable ladder

The project is worth 30% of your grade in the 4-credit version of the course, and completing it is what earns the six sigma black belt certification. You build it across the term in ten steps. Each rung is graded on its own and feeds the next; together they become the final report and the poster. Every step is submitted on Canvas — see there for the current due dates.

1. Team & topic

   Form a team of 3–5 and commit to one process you can actually observe or get data about. Pick from the catalog above or bring your own — the constraint is access, not ambition.

   Your project team and your homework team are separate Canvas groups. Both were seeded from your Starter Team, so they line up right now, but they do not have to stay the same — talk to the teaching team if your project team ends up different from your homework team.

2. Your dataset

   Deliver the actual data — experimental, observed, simulated, or found — with a codebook that says what every column means, what its units are, and where each number came from. A plan to find a dataset later does not count.

   Details What to submit by project type (Experimental · Simulation · Monitoring) ▾

   Submission Requirements by Project Type

   Depending on whether your project is an experimental study, a simulation project, or a monitoring project, the structure of your dataset submission is different. In all cases, submit up to 10 variables (or parameters) along with a complete codebook.

   In All Cases: Provide a Codebook Table

   A codebook is a clear reference table describing every variable in your data. Each variable must have its own row in the table, with three columns:

   Variable / Column Name	Unit of Measurement	Definition & Role
   cure_time	minutes	Predictor: duration the adhesive bond is cured at 65°C.
   shear_strength	MPa	Outcome: peak mechanical shear stress measured before joint failure.
   operator_id	category (A, B, C)	Covariate / factor: technician who prepared and assembled the test coupon.

   EXPERIMENT

   1 · Experimental Study (Designed Test / Intervention)

   State what variables you will collect: You have not run your experiments yet, so you have not collected the data — you are reporting what you will collect (planned variables, maximum 10 variables). Use “we will collect” phrasing.

   • Outcomes vs. Predictors: Clearly distinguish what outcomes you are measuring and what predictors you are testing. Predictors are the input factors, treatments, or operating conditions you think will affect the outcomes.
   • Measurement Scales & Units: There are many different valid measurement scales you can use:
     • Real physical quantities: grams, meters, seconds, degrees Celsius, beats per minute, etc.
     • Likert scales: 1 to 5 (or 1 to 7) ratings for satisfaction, perceived difficulty, or product preference. Likert scales are a completely valid measurement option!
     • Counts: discrete integers like 0, 1, 2, 3, 4, 5, 6 (e.g. defect counts, event occurrences).
     • Percentages: 0% to 100% (e.g. process yield, inspection accuracy, defect rate).
   • Design & Recording: Specify your planned sample size per condition (number of observations per treatment cell), who will record the observations, on what instrument/sheet, and in what order.

   SIMULATION

   2 · Simulation Project (Published Parameters & Models)

   Your dataset becomes parameters: In a simulation study, your data deliverable consists of the statistical parameters that drive your model. You must supply at least 10 parameters or components.

   • Acceptable Parameter Examples: Relevant reliability and process statistics for each component or sub-process, such as:
     • Failure rate (λ)
     • Mean time to failure (MTTF or MTBF)
     • Probability of failure at a specific time t for a specific sample size n (e.g. P(Fail ≤ t) or reliability R(t))
     • Having any of those three metrics for your components works great. Additional parameters include standard errors and distribution shapes (e.g. Weibull β and η, Exponential, Normal).
   • Full Bibliographic Citations: Every single parameter must come from published literature, manufacturer datasheets, or certified reliability registries with a full citation. No invented numbers or synthetic placeholders.

   MONITORING

   3 · Monitoring Project (Process Observation Over Time)

   Describe your data source: Clearly identify where the data comes from and state whether it is live or already collected / historical.

   • Source Examples: A live stream/API (e.g. real-time traffic feed, live server telemetry, live weather station) versus historical records (e.g. historical air quality records from the EPA AQS API, municipal open data portals).
   • Outcome Variables (Max 10): Identify what outcome variable(s) you are studying. For a high-quality monitoring project, you only need 4 or 5 variables in total to run an outstanding analysis.
   • Observation Over Time: Quality monitoring projects track behavior dynamically. Define both:
     • Unit of observation: The specific entity or location being measured (e.g. intersection, sensor station, production machine).
     • Unit of time: The sampling cadence or time interval (e.g. minute-by-minute, hourly, daily, per shift).

3. Project charter

   One page stating the problem, the scope, the metric you will move, and the target. If your team disagrees about the project, the charter is where it shows up.

   Instructions

   Project Charter

   Estimated time: about 1 hour.

   The Council for Six Sigma Certification requires Six Sigma Black Belt projects to include a Project Charter: a single-page document that summarizes your project.

   A project charter is the living document for an improvement team that outlines the problem, target, and boundaries of a process improvement effort. It is good practice as a systems engineer and great for communicating and clarifying the exact expectations and deliverables for your client.

   • Fill out the Project Charter form (template here) to clarify your project.
     • For the timeline, show real dates from this term ahead.
     • Goals and deliverables must be realistic for your project.
     • Be specific.

   Target audience: think of this document as what you might submit to a top manager (e.g., a Chief Technical Officer) or a funding agency to obtain support for your project. What information would a CTO or funding agency want to see?

   Note: roles are made by you, e.g. Project Lead, Project Manager, Visualization, Modeler. "Sponsor" only applies if your team has an actual sponsoring organization giving you data.

   To submit

   Upload a .docx file containing your completed template.

   Charter template

4. VOC tree & SIPOC diagram

   Two halves of the same framing move, now one submission. The voice-of-customer tree turns what the customer says into a measurable requirement; the SIPOC bounds the process around it — suppliers, inputs, process, outputs, customers — so you know what is inside your project and what is somebody else's problem.

   Instructions

   Voice of the Customer tree

   Turn what the customer says into measurable requirements. Start from the VOC tree template and customize it to your project; keep an editable version for your final report, where it is integrated with a step-by-step description in the main text.

   SIPOC Diagram

   Estimated time: about 1 hour.

   A SIPOC diagram is a visual tool used in process improvement projects to map out the key elements of a process. It stands for Suppliers, Inputs, Process, Outputs, and Customers. It provides a high-level overview of the process and helps stakeholders understand its scope and boundaries. It includes:

   • Suppliers: who or what provides the inputs to the process?
   • Inputs: what materials, information, or resources are needed for the process?
   • Process: the series of steps or activities that transform inputs into outputs.
   • Outputs: what is produced or delivered by the process?
   • Customers: who receives the outputs or benefits from the process?

   Fill out this SIPOC diagram template, customizing it to match your research project. You may use Google Drawings, PowerPoint, Draw.io, Mermaid, or any other valid tool for visualizing the diagram.

   Your SIPOC diagram should:

   • Feature the process being studied (e.g. making coffee, making semiconductors, delivering medical care), not the process of studying it. "Make regression model" or "write report" should not be in your SIPOC diagram.
   • Be visually appealing, with clear and effective use of color.
   • Be readable; it should not require your reader to zoom in 1000%.

   Target audience: think of this document as what you might submit to a top manager (e.g., a Chief Technical Officer) or a funding agency to obtain support for your project. What information would a CTO or funding agency want to see?

   To submit

   Upload a .docx file containing your diagram(s).

   Note: keep an editable version of the chart for yourself. You will need it for your final report.

   VOC tree template SIPOC diagram template

5. Process map & literature review

   Open the middle of the SIPOC into the real sequence of steps, decisions, and handoffs — including the rework loops nobody puts on the official diagram — and, in the same submission, establish what is already known about this failure mode or process. Cite what you use; the point is to avoid re-deriving a published result.

   Instructions

   Process Map

   Open the middle of the SIPOC into the real sequence of steps, decisions, and handoffs. Start from the process map template and customize it to your project; keep an editable version for your final report, where it is integrated with a step-by-step description in the main text.

   Literature Review

   Estimated time: 1–3 hours.

   Background / literature review / definition of project.

   • Task: summarize relevant prior scholarly research and professional practice on your topic, drawing on journal articles and reports from individual organizations. We do not expect a full literature review, but be sure to include at least 20 sources.
   • Focus on recent work (which often includes synopses of early work). Your review will give you good ideas for your methods, model design, and analysis. Cite all sources appropriately in the reference list.
   • E.g. for market or industry topics, describe the essential features of that market or industry. For policy questions, describe the legislation and its background.
   • References: at least 20 APA-formatted references.
   • In-text citations: every reference must be cited in the text at least once.
   • Types of sources: a mix of scholarly journal articles, reports, and, where necessary, news articles or other sources.
   • Goal: must directly connect to your research question. Helpful ways to write a literature review:
     1. What are the alternative explanations for your question?
     2. What is the background on the product, what are its known problems, and what is the treatment under study?
     3. Compile a table of parameters from past studies, summarizing their findings. What are the common findings? What was not well assessed or measured, or could not be, and how does your study fix that?
     4. Use your SIPOC diagram and/or process map, and explain each portion of it using background literature. What is not yet well understood?

   Target audience: think of this document as what you might submit to a top manager (e.g., a Chief Technical Officer) or a funding agency to obtain support for your project. What information would a CTO or funding agency want to see?

   To submit

   Upload a .docx file containing your literature review.

   Note: you may be assigned to peer review two other teams' literature reviews. Take it as a chance to see what other teams are doing. Compliment-sandwich feedback works well: three comments, two good and one area for improvement.

   Process map template

6. Research design

   Say exactly how you will get evidence: what you measure, how often, under what conditions, and what result would change your mind about the cause. Any experiment you intend to run on people gets vetted by an instructor before you run it.

   Instructions

   Research Design

   Estimated time: about 1 hour.

   Fill out the research design form (template here) to design your project. Key components:

   • Question: one concrete, testable research question.
   • Measures: 2–3 specific measures you will use.
   • Data sources: where you will source these data from; simulated data is fine as long as it comes from real parameters from a previous study.
   • Methods: the specific methods and techniques from the course you will use to answer your research question.

   Target audience: think of this document as what you might submit to a top manager (e.g., a Chief Technical Officer) or a funding agency to obtain support for your project. What information would a CTO or funding agency want to see?

   Formatting requirements

   • Use in-text citations.
   • Include a references section listing any works cited. Use a proper and consistent referencing style; if unsure, ask the instructors or consult the referencing styles used in journals.
   • Avoid footnotes; they are only for supplemental material not strictly needed in the main text.

   Other notes

   • Your analysis could lead to empirical work, hypothesis testing, experiment design, or systems optimization/control. It is fine to mention developing these implications as an end product of your work.
   • For DL students and students with work experience, this project is a great opportunity to talk with your colleagues or managers about how the course project could improve a product or service your company provides. Previous DL students (and their managers) responded very positively to initiating a project closely related to their work. The instructors are ready to discuss project ideas to get you on the right track and form teams early.

   To submit

   Upload a .docx file containing your completed template.

   Research design template

7. Preliminary results & financial impacts

   First pass at the analysis — the charts, the fitted models, the intervals — reported together with what the improvement is worth in money: cost of poor quality now, cost after, and what the fix would take. Early enough that a broken measurement system is still fixable.

   Instructions

   Preliminary Results

   Estimated time: 3+ hours.

   Summarize your preliminary methods and results. Your response should include:

   • Measures [100–150 words]: what are you measuring?
   • Assumptions [~100 words]: adequately describe your assumptions. Your topic is undoubtedly complex, so you will probably simplify it using assumptions, taking some things as given; explain and justify the choices you make.
   • Model / approach [100–300 words, minimum]: a large portion adequately describing your model, approach, or techniques, and exactly how and why they help you answer your research question. (Detailed, not "we made a regression model. Next!")
   • Results [300 words, minimum]: summarize the numbers generated by your statistical analyses.
   • Key numbers in text: report the important numbers from your analysis in the text, with attention to statistical significance and confidence intervals.
   • Table / figure of results: report your results transparently in a table and/or figure. If you made a regression model, report the regression table, and so on.
   • Must show results: your response must actually show results, not a plan to do something in the future. That was the Research Design; this is the Results. No credit is given without results.
   • Budget your time: this takes real time and effort. Reach out early to the instructor and/or TA with questions.

   To submit: upload a .docx file containing your response. Keep an editable version for yourself; you will need it for your final report.

   Financial Impacts

   Estimated time: about 1 hour.

   The Council for Six Sigma Certification expects all Six Sigma Black Belt applicants to be able to meaningfully estimate the financial value produced by their engineering work for a client. Demonstrate the potential financial impacts based on your project analysis.

   Draft a 300–500 word "Financial Impacts" section of your Six Sigma report's Discussion, clarifying the specific financial impacts based on your results. It should include:

   • Reasonable cost estimates based on specific documents or online resources. (Google it!)
   • In-text citations and bibliographic references for any cost estimates or external data referenced.
   • A cost comparison built from your results: how much X would cost before vs. after implementing a specific change proposed from your results.
   • Realistic estimated benefits from your proposed change.
   • Confidence intervals for your financial impact estimate(s). Have a plan for calculating them. Often there is no existing formula for a standard error in your case; you may need bootstrapping or simulation to create confidence intervals or uncertainty estimates. These estimates must be real and valid.

   Your response should be evidence-based. Do not make wild claims.

   See this example using the Hot Springs data: ZZ_financial_analysis.R.

   Target audience: think of this document as what you might submit to a top manager (e.g., a Chief Technical Officer) or a funding agency to obtain support for your project. What information would a CTO or funding agency want to see?

   To submit: upload a .docx file containing your response. One student can submit on behalf of the whole group.

   Financial analysis example

8. Rough draft

   The whole report assembled end to end while there is still time to fix it. Complete enough to get real feedback on — gaps named as gaps, not quietly left out.

   Instructions

   Rough Draft

   Estimated time: 3+ hours.

   Create a rough draft of your Six Sigma report, synthesizing the diagrams and analyses you created previously. Submit:

   • A single-spaced report-in-progress with 1800–2100 words of new content (about 5–6 pages).
   • The report updates the status of your project: what you have done, what remains and the timeline, any major problems and strategies to attack them, and references used.
   • Word limits are a minimum, not a maximum. If you need more words, good, go for it.

   You can follow this general structure:

   • Introduction
     • Project Charter (template; already done, not in the word count)
     • Background and project motivation [100–150 words]
     • Project objectives and expected contributions [100–150 words]
   • Literature review
     • SIPOC Diagram (template; not in the word count)
     • Process Map (template; not in the word count)
     • Voice of the Customer Tree (template; not in the word count)
     • Background literature [200–300 words]
   • Methodology
     • Research Design (template; already done, not in the word count)
     • Data [100–150 words]
     • Measures [100–150 words]
     • Model / approach [100–300 words]
   • Preliminary results [300+ words]
   • Discussion [300 words]
     • Summary of findings [~100 words]
     • Financial impacts [200 words]
   • References cited (not in the word count)

   Target audience: think of this document as what you might submit to a top manager (e.g., a Chief Technical Officer) or a funding agency to obtain support for your project. What information would a CTO or funding agency want to see?

   To submit

   Upload a .docx file containing your draft.

9. Final report

   The whole arc in one document — define, measure, analyze, improve, control — ending in a recommendation somebody could actually act on. This is the deliverable the black belt certification is judged against, so it is expected to be thorough and complete.

   Instructions

   Final Project Report

   Estimated time: 5–10 hours.

   Create a final report, with all code included in the appendix. The report should meet each of the following requirements (R1–R5), weighted as indicated.

   R1 · Introduction (10%)

   • Present your topic (any current quality or reliability issue with a product, service, or integrated system) and argue for its importance to a meaningful group of people or clientele (e.g. top business managers, a quality team leader, policy makers). You can introduce a broad topic, but your project must focus on one narrow aspect that can be analyzed in depth in one semester.
   • Integrate previous diagrams here: Project Charter (template; already done).

   R2 · Background / literature review / definition of project (15%)

   • Summarize relevant prior scholarly research and professional practice on your topic, drawing on journal articles and reports from individual organizations. Not a full literature review, but at least 20 sources. Focus on recent work. Cite all sources appropriately in the reference list.
   • E.g. for market or industry topics, describe the essential features of that market or industry; for policy questions, describe the legislation and its background.
   • Integrate previous diagrams here, each with a detailed, step-by-step description in the main text: SIPOC Diagram (template), Process Map (template), Voice of the Customer Tree (template).
   • Reference list and in-text citations (at the end of the paper).

   R3 · Measure and design (40%)

   • Summarize your methods, describing your data and variables, assumptions, specific models and techniques, and exactly how and why they help you answer your research question. Includes: Research Design form (template; already done), Data, Measures, Assumptions (your topic is complex, so you will simplify it using assumptions; explain and justify the choices you make), Model / approach / techniques.

   R4 · Analysis, recommendations, and conclusions (25%)

   • Report your results clearly and transparently, in accurate but easily understandable terms. Explain your analysis and modeling, key findings, any limitations, and the next steps needed to solve the problem or answer the question. Includes: Results (including the financial impacts analysis), Discussion, Recommendations and conclusions.
   • Implementation Plan Table (template), with a detailed, step-by-step description in the main text.
   • Control Plan Diagram (template), with a detailed, step-by-step description in the main text.

   R5 · Format and writing (10%)

   • Basic formatting (5%): 18–40 double-spaced pages, excluding pages of graphs, tables, reference list, etc. Include your code in an appendix.
   • Quality of writing (5%): good grammar, spelling, and sentence construction, but also clear structure and organization, flow, and fair presentation (strengths and weaknesses). Your target audience is a practitioner with at least graduate-level training, not necessarily a hotshot theorist. Assume a firm advisor trained in systems engineering will read your report and disseminate your results to end users.

   Extra credit (up to 8%, where relevant)

   • Quality of team effort.

   Reminders

   When finishing your final report, lightly check it for the following; these are standard expectations in a formal Six Sigma report, but easy to forget in a busy end of term.

   • Your Project Charter fits on one page (it does not have to be double spaced).
   • Cite your tables, figures, and equations in the main text: "As shown in Table 1..." or "Equation 1 shows...".
   • Discuss your tables and figures in the text. Walk the reader through what the figure shows: each step of your monitoring plan; which bar was highest vs. lowest. You need not describe every part, but it is your job to interpret the plot for the reader.
   • Reuse your Project Charter's content in the introduction, if helpful. Many charters hold excellent stats and crisp summaries; technically the charter is a table, a stand-alone item, so it does not count as main text and a reader may skip it. You are welcome to reuse its wording and stats in your introduction.
   • Good reports = good writing. Use transitions and summary statements at the start of new sections, headers and subheadings, and bolding or italics if helpful. Example transition: "Next, we introduce our approach for statistical analysis..." Example summary statement: "This included (1) modeling the lifespan distribution, (2) estimating the lifespan of a cookie, (3) modeling the cost of a cookie, and (4) projecting the cost over time. First, to model the lifespan distribution, we..."
   • Appendices go at the end of the report, after the bibliography (particularly code; your reader will fall asleep reading code, so it goes at the very end). Label them Figure A1, A2, Table A1, Equation A1, Figure B1, and so on.
   • Include your commented code. Transparency is key to showing the client that you did in fact do the analysis.

   To submit

   Upload a .docx or PDF file containing your report. It must include an appendix with your complete code.

   Implementation plan template Control plan template

10. Poster presentation

    The report compressed to what a stranger can absorb standing up: the problem, the evidence, the fix, and the number that proves it worked. Presented live to the class.

    Instructions

    Poster Presentation

    Estimated time: 3+ hours.

    Create and present an academic poster of your Six Sigma project at the public poster session (see Canvas for the time and place).

    Why a poster? This format helps you (1) explain your project concisely to future employers and (2) produce a concise, beautiful poster for your professional portfolio.

    Requirements

    • Contents: (1) research question, (2) data, (3) method, (4) results, and (5) discussion/implications.
      • Brief bullet points succinctly explaining each section; avoid walls of text.
      • Two ggplot visuals describing your findings: informative, visually appealing, and well labeled. E.g. a fault tree visual, coefficient plot, statistical simulations, process overview, or any technique from the course.
      • Clearly stated quantities of interest (single numbers that answer your research question).
      • Confidence intervals for any quantities of interest, e.g. simulations, probabilities, beta coefficients, predictions.
    • How do I make a poster? Review Cornell Library's guide to making posters.
    • Templates: you can make your poster in PowerPoint / Google Slides, LaTeX, among other methods.
    • Dimensions: aim for 42 inches wide by 36 inches tall.
    • Deliverables: (1) bring your PDF and (2) submit your poster PDF on Canvas.

    Poster session details

    • Rotation: each team's poster is mounted on a board and you succinctly summarize your work for attendees. For the first half of the session, half of the teams present while the other half look at posters; for the second half, the teams swap.
    • Informal presentations: be prepared to describe your work in clear, everyday language as attendees view your poster. A good chance to work on your elevator-speech skills. Team members should take turns.
    • Attendance: mandatory for Black Belt students, but everyone is welcome. Come support your peers.
    • DL students: if your entire team is DL, pre-record a short 2-minute video of your poster presentation (a Zoom recording is fine) and submit it as a comment attachment on the assignment; DL posters are shared and displayed at the session. If some team members are not DL, they present the poster, so contribute to the poster design instead.

    Past posters and examples

    • Past posters from this class (on Canvas, for enrolled students — see the card below)
    • Academic poster examples: Cookstove Emissions, Stroke Prediction, Reddit Post Popularity Prediction, other health sciences examples. Further example posters from past students are posted on Canvas for enrolled students.
    • Shiny user showcase: shiny.posit.co/r/gallery

    To submit

    Upload a .pdf file containing your poster. If DL, also include a working share link to a short video presentation.

    Bonus: if you make a Shiny app dashboard, you are eligible for a bonus on this assignment.

    Poster template