1 · Introductions
Wednesday, August 26, 2026
Due before meeting 2 · Tuesday September 1
Research interest survey, on Canvas, by 11:59 pm. Under an hour. Most answers are two or three sentences.
This replaces the one-page memo described in class on August 26. Do not research whether your data exist. An idea with no data plan is fine at this stage. Two of the questions, what you bring and where this fits your plans, are how teams get matched.
Session A · 2:05–3:15
The arc
AEM 6991 produces a plan; AEM 6992 executes it. The deliverable in December is a proposal somebody could hand to a team in January.
AI ground rules
Permitted by default, disclosure mandatory, verification graded. From meeting 12 you work with real data through an assistant, and what is graded is the questions you ask and the checks you run — never code you wrote.
What this is, and is not
You are not writing a paper for a journal — there is no gap to fill and no contribution to claim. You are building the analysis a competent team would need before making a decision that costs real money, and the standard for “competent” is the same standard a referee would apply. Rigor and genre are different axes: the rigor bar does not move, the genre does.
This is also why the literature matters without the project being a literature project. You read to find out whether your question is already answered, and how people have answered questions like it.
For context, MPS AEM class of 2025: 40% into finance, 21% marketing and sales, 15% general management, 13% consulting, 6% data and analytics. Essentially none into research.
Session B · 3:30–4:30
What you already have access to
Before assuming data must be bought or scraped: the Cornell library’s licensed holdings, and Dewey, a commercial data marketplace the Johnson College of Business subscribes to and which you can use at no cost. Teams routinely propose buying data the College already has, or drop a good question because they assumed the data was unreachable.
Register at app.deweydata.io with your Cornell email, in week 1. Your account sits pending until an administrator approves it, so do not leave it until you need the data.
If you are stuck, there is a numbered list of leads under Resources — potential project ideas. Treat it as a way to get unstuck, not a menu. Nobody has ever written a good capstone from somebody else’s idea list.
Working backwards from a dataset is legitimate — it is route 5 — provided you do it honestly: what could this answer, which of those would anybody act on, and what does it not contain that you would need? The question still has to survive the filter, and you are still allowed to walk away from the data. Route 5 done badly is anti-pattern 2.
Your interests
Teams do not exist yet. They are formed from what you tell each other this afternoon and from the research interest survey you fill in this week.
Round one. In circles of five, one at a time: you say what you would like to explore and why it interests you. The person on your left says it back in one sentence. You confirm it, or fix it in one sentence. Everyone speaks once and summarizes once.
The summary is the exercise, not the speech. Your job while somebody talks is not to plan your own turn. Most summaries come back subtly wrong, and the correction is the moment a vague interest gets sharper — which is the whole course in miniature. If somebody else cannot restate your claim, it is not clear yet.
Round two. New groups. Describe your idea again, to people who have not heard it, and the group names the obstacles. Not you — the owner of an idea is invested in it and has usually been avoiding the obvious problem, and somebody hearing it cold names it in ten seconds. Write them down.
Report back. Each group tells the room two or three: what the person wants to explore, and the biggest obstacle the group found. You never report your own.
Every idea goes on the board, because seeing them all at once is how overlaps become visible.
These are not the teams.
Where to look for data
Before assuming data must be bought or scraped: the Cornell library’s licensed holdings, and Dewey, a commercial data marketplace the Johnson College of Business subscribes to and which you can use at no cost. Teams routinely propose buying data the College already has, or drop a good question because they assumed the data was unreachable.
Register at app.deweydata.io with your Cornell email, tonight. Your account sits pending until an administrator approves it, so do not leave it until you need the data.
Your team’s output is submitted before you leave.