Interpreting a Scatterplot of Delivery Times
A logistics dashboard plots daily delivery time (hours) versus number of stops for 30 days. The points form an upward trend but with wide spread. Two clusters stand out: weekdays cluster around 6–8 hours at 40–50 stops, while weekends show 9–11 hours at only 30–35 stops. A manager concludes 'more stops cause longer days' and proposes capping stops at 35 to shorten delivery times. A note on the chart says weekend routes include a mandatory 2-hour depot audit before departure.
1. Does the scatterplot support the manager's conclusion?
Read the worked answer
No. The two clusters confound the relationship. Weekends have fewer stops but longer days, which contradicts a simple 'more stops = longer day' reading. The stated 2-hour weekend audit explains much of the weekend cluster's elevation. Pooling the two groups creates a misleading trend line because a hidden variable (day type / audit) drives both stop count and total time.
2. What is a sounder recommendation?
Read the worked answer
Analyze weekdays and weekends separately. Within weekdays, the stop-count relationship may hold and inform a cap. On weekends, the fixed 2-hour audit dominates; capping stops at 35 would not shorten weekend days and could hurt coverage. Recommend separating the data, then testing a stop cap only on the weekday cluster where the trend is real.