Mini-project
Banking & Fintech
ATM cash demand forecast
Forecast daily cash withdrawals so ATMs neither run dry nor hold idle cash.
The brief
A bank's cash team loads ATMs on a fixed schedule; some run out on salary days while others sit full. Forecast daily withdrawals for a set of ATMs, with weekly cycles, month-end salary days and festivals, and suggest how much cash to load. Good looks like: a forecast that beats a seasonal baseline across many ATMs, and a loading plan with its cost.
What you will build
- Prepare daily withdrawal series for many ATMs
- Add calendar features: weekdays, month-end, festivals
- Compare seasonal-naive, ARIMA and gradient-boosted forecasts
- Back-test over several months
- Turn forecasts into a loading plan with stock-out risk
Free, open data to build it with
- NN5 daily ATM withdrawals (Monash Time Series Forecasting Repository) ↗ Monash University
Check each dataset's licence before you publish your work. Your mentor gives your team its own version of the brief (data, scope or region), so every team's project is its own.