🚕 Haggis Hopper

Business Insights Dashboard

Glasgow taxi trip data for February 2024 — demand patterns, pricing, revenue, and a short-horizon forecast, in one view.

Demand patterns

When trips actually happen — by hour of day and day of week.

Trips by hour of day

24-hour view across the full month

Trips by day of week

Totals across all Februaries in the sample

Rush hour heatmap

Trip volume by hour and day of week — darker means busier. This is the clearest single view of when to have drivers on the road.

Fewer trips
More trips

Pricing analysis

How fares move through the day, and their overall distribution.

Average fare by hour

£ per trip, by pickup hour

Fare distribution

Trip count by fare band

Revenue by area

Pickup area ranked by trip volume. Named areas are sourced from prior Tableau analysis of this same dataset; the rest fall under a generic "Glasgow" grouping in that same source, shown alongside their postcode district for precision.

Area Trips Revenue Avg fare Avg duration Revenue/km

Fare & duration by hour and day

Richer than the hour-only views above — the duration heatmap in particular is a traffic-congestion signal: longer trip times at fixed distances during commute hours, not just more trips happening then.

Median fare

By hour and day of week

Average duration

Darker = longer trips at that hour — watch for rush-hour congestion

Where demand happens, by hour

Which areas are busy at which times — a geographic view the day-of-week heatmap above can't show on its own.

Pickups by area & hour

Where trips start

Drop-offs by area & hour

Where trips end — compare against pickups above to spot commute flows (e.g. an evening pickup hotspot paired with a same-hour drop-off hotspot elsewhere suggests a real route, not just two independently busy areas)

Origin → destination flows

Trip counts between the 12 busiest areas — actual routes, not just per-area totals. Rows are pickup, columns are drop-off.

Passenger analysis

How party size relates to fare and tipping behaviour.

Trips by passenger count

Fare & tip by party size

PassengersTripsAvg fareAvg tip

Correlations

How trip characteristics move together. Blue = positive correlation, red = negative, gray = little relationship.

Notable relationships

    Pricing outliers & anomalies

    Statistical outliers in distance/duration/fare, plus trips whose fare-per-km rate looks off relative to the rest of the book.

    Data quality:

    Fare-per-km anomalies

    Trips charging >3x or <0.33x the median rate
    TimeDistanceFareRatevs median

    Most extreme trips

    By distance — worth a manual look
    TimeDistanceDurationFare

    Time series forecast

    Daily trip volume with a 7-day forward extrapolation.

    Method:

    Pricing model

    Predicting fare from trip characteristics (distance, duration, passengers, hour) — a different question from the demand forecast above: this predicts price for a given trip, not future trip volume.

    Feature importance

    Random Forest — which trip characteristics drive the fare prediction

    Recommendations

    Data-driven, not generic — each one ties back to a specific number above.