Assignments Due
Week 12 checkpoints should be submitted by Monday, 04/02/2018 at 11:59 PM PDT.
Compose
Final: Design (33%)
Your client is proposing a data science investment at next month’s board meeting, and she directs your data science team to prepare a design memo and presentation to persuade the board to greenlight the concept. Time to stretch it out, munch on some user stories, and scrum your way to the finish line!
The Original Final Prompt
You’ll select a domain of interest and explore the current state of data science in that domain. What are analysts currently doing with data, and what do you think they could or should be doing? You’ll identify what types of difficult and important questions confront decision makers in that domain and what data they may now be using to address those questions.
Using that as background, you’ll select one question you believe data scientists in this domain should explore and then detail how you would design a study, service, or product to address that question with data science in a manner that would aid decision makers.
Tell us about how you would acquire the needed data and, if appropriate, the ethical and legal issues of using these data. Propose how you would design and conduct the study, convey your findings to relevant audiences, and persuade decision makers to act differently on the basis of your (prospective or retrospective) findings.
Outline potential risks for this project, such as data access or data quality challenges, as well as organizational and decision impediments to action, and describe your contingencies given these risks.
Last, propose future questions you could address with findings from this project.
Your Revised Final Prompt
Above you see a fine and dandy final proposal prompt. Yet after three months of practice you’re skeptical. The original prompt (let’s call it the OP) sounds a little old-fashioned, a little too “researchy”, with too much BI and not enough DS. You’d like to do something more interesting, and in your own style. You’d like to address a prompt better tailored to your big idea.
That prompt is just too vague, isn’t it? You think you can do better. You can do better!
Your initial task will be to deconstruct the OP by identifying the features asked for and by organizing these according to the invisible ink you suspect is underlying the OP. Reveal its hidden logic and consider why and whether the ingredients described there add up to a good recipe for data science.
Second, rewrite the OP according to your own vision of data science quality. Do this by making intentional decisions about elements of invisible ink such as audience, armature, and story structure. After laying out the ingredients of the OP, decide what to keep, cut, rearrange, and add. For guidance consider lectures and readings you have found interesting, reference materials like the DSM template, your own goals and ambitions, and any real paradigms you may want to emulate.
Conduct this exercise in a team meeting with your group where you argue together about what will make for a great data science concept, keeping the team leader’s proposal in mind as a strong opening salvo.7 Together, literally rewrite the prompt as a directive you will give yourselves and then execute toward completion of the final.
Let this process serve as a constitutional convention for the team. You will be asked to discuss your new prompt in consultation with your instructor in live session, and to include it as an appendix in your write-up. Maintain the imperative mood of the original prompt: give yourselves marching orders for designing a strong concept!
Your final will consist of a write-up, which will include your revised prompt as an appendix, as well as a presentation. Consider role-playing each document itself. What would your revised prompt look like as a one page memo from your client detailing the requirements of your data science proposal? How might you format the write-up for submission as a corporate white paper or as a grant application for government or foundation funding? How would your presentation fare before a keynote audience at a trade show, in front of a board of managers, or facing publicly on social media? It will be more fun for us and interesting for you if you simulate a realistic setting for your product, which means avoiding treating the instructor or the class as the audience for your work.
A sample of an effective revised prompt from students in a previous cohort can be found in Appendix A of the RecipeMate proposal. Just the front and back matter, not the body, are included as examples, but please be clear that you need to craft your own strategy. It is not strictly necessary to role-play the revised prompt, but it is a recommended method for ensuring that you make a strong client connection to your work.
How to submit
You will begin your final to be turned in at the end of the term, but your work-in-progress will be noted by the instructor the morning before class. You will not be required to prepare your final project as a GitHub repository. Instead you will use a Google doc to make collaboration a bit easier and to allow you to focus on content and not syntax.
Start by making a copy of the write-up template.8 For Week 12 you will be responsible only for writing your revised prompt as a one page memo and including it as Appendix A in your copy of the template.
Store your version of the template in this bDrive folder. You will fill it out over the next two weeks.