Outlier Treatment for Revenue Reporting | Dataford Interview Questions

Your question is Outlier Treatment for Revenue Reporting. Take a moment with it on the right.

Talk me through your thinking if you like. When you're confident, submit your answer and I'll grade it like a real screen (7/10 or better passes).

Business Context

StreamCart produces a daily revenue-per-order report used by finance and operations. A recent pipeline run showed a sharp jump in average order value, and the team suspects a few extreme transactions are distorting the metric.

Problem Statement

You need to evaluate how outliers should be handled in the reporting pipeline and quantify the impact of different summary statistics. Use the sample of daily order values below to identify outliers with the IQR rule, compare the mean before and after excluding outliers, and compute a 95% confidence interval for the cleaned mean.

Given Data

A sample of 15 order values (USD) from one day:

Order Values
42
45
47
48
49
50
51
52
53
54
55
56
58
60
220

Assume the reporting team currently publishes the arithmetic mean as the headline KPI.

Requirements

Sort the data and compute $Q_1$ , $Q_3$ , and $IQR$
Use the 1.5 $\times$ IQR rule to identify outliers
Calculate the mean with all observations included
Calculate the mean after excluding detected outliers
Compute the sample standard deviation of the cleaned data
Construct a 95% confidence interval for the cleaned mean using the t-distribution
Recommend how the pipeline should handle outliers for reporting and monitoring

Assumptions

This sample is representative of one reporting day
The cleaned observations are approximately independent
After removing clear outliers, the mean can be summarized with a t-based confidence interval
The goal is reporting robustness, not fraud detection or root-cause diagnosis