Week 2 Tutorial: Dispatch lab
A hands-on session reverse-engineering a real delivery app's ETA and batching behaviour from the outside
Graded checkpoint, checked live at the start of the session · Weight: 2%
Before the session#
Build a cart in a delivery app and step through to checkout — no need to pay — logging the ETA it shows at three or more points: cart added, address entered, checkout reached, and any jump in between. A phone screenshot with a timestamp is enough; a spreadsheet row per reading is better, since week 3 reuses this log’s format.
In the session#
Working in pairs with your logged readings, reconstruct where the shown ETA moved and why — a batching assumption changing as the cart grows, a recalculated distance once an address is entered, a jump between checkout steps that has no obvious cause. Groups compare notes across apps: does Menulog’s ETA move the same way Uber Eats’ does?
Afterwards#
Your ETA log becomes raw material for week 3’s case-study reading — keep it.
The spec
What you bring to the session, checked live: a checkable minimum plus one item that's judged by discussion, not marked for prose.
- at least three timestamped ETA readings for the same cart, logged from a named app you can actually use here (Uber Eats, DoorDash, or Menulog)
- one moment the ETA changed is marked, with a specific proposed reason
- you have named one design decision from the lecture reading that a rider would experience differently than an engineer intended
Teaching team
- Chajie Zhou — convenor