This project started from a simple operational question: how can a mobility business plan availability with more confidence instead of reacting too late to peaks and gaps in demand?
Objective
The idea was to build a forecasting workflow that could help with:
- vehicle allocation across locations
- expected utilisation by day and week
- identifying demand patterns around seasonality and local events
- supporting better operational and pricing decisions
What I built
I designed a small pipeline that prepares historical bookings, cleans operational gaps, and creates features for trend, weekday behaviour, lead time, and seasonal effects.
From there, the workflow produces forecasts at a level that is useful for operations rather than just for model accuracy. The point was not to create the most complex model possible, but to generate outputs that a business could actually use to plan inventory, staffing, and maintenance windows.
Why this matters
Forecasting becomes valuable when it improves decisions. In this case, the practical use is clearer visibility into expected demand and earlier signals when supply or pricing strategy may need to change.
Outcome
This case study reflects how I approach analytical problems: combine clean data preparation with simple, explainable modelling, then turn the results into something operational teams can use without needing a data science background.