Visualize Trip Duration, Fees, and Survey Outcomes

Quick Overview

Design a visual analysis for trip duration, fees, binary survey outcomes, and a fourth segment. Cover distributions, scatter or density plots, faceting, overplotting, and response bias.

Visualize Trip Duration, Fees, and Survey Outcomes

Company: Amazon

Role: Business Intelligence Engineer

Category: Analytics & Experimentation

Difficulty: easy

Interview Round: Technical Screen

A mobility service tracks trip duration, trip fee, and a post-trip survey outcome recorded as `good` or `bad`. Design a visual analysis that progresses from one metric to two, three, and four variables. Explain which charts you would use, what each reveals, and how you would avoid misleading encodings or overplotting. For the fourth variable, choose one useful dimension such as city, vehicle type, or time period and state why it belongs in the analysis. ### Constraints & Assumptions - Duration and fee are positive continuous variables with possible right skew and outliers. - Survey outcome is binary and may be missing nonrandomly. - The audience needs both a high-level summary and the ability to diagnose segments. - Do not encode several continuous variables only by bubble area without explaining perceptual limitations. ### Clarifying Questions to Ask - Is the goal operational monitoring, causal diagnosis, or executive reporting? - What is the observation grain: trip, rider, route, or daily aggregate? - Are fee and duration comparable across cities and vehicle types? - What fraction of trips receive a survey response? ```hint Add channels carefully Start with distributions, then a duration-fee scatterplot. Add the binary result through color or separate panels and reserve faceting for the fourth variable. ``` ### What a Strong Answer Covers - Appropriate single-metric summaries rather than a context-free large number. - A scatterplot or density view for duration and fee with skew and overplotting controls. - A clear encoding for binary survey outcome and a comparison of conditional rates. - A justified fourth dimension through facets, small multiples, or stratification. - Missing-response bias, aggregation grain, and accessible color choices. ### Follow-up Questions - How would you separate the effect of trip duration from city mix? - What chart works when there are millions of trips? - How would you communicate survey nonresponse to an executive audience?

Quick Answer: Design a visual analysis for trip duration, fees, binary survey outcomes, and a fourth segment. Cover distributions, scatter or density plots, faceting, overplotting, and response bias.

|Home/Analytics & Experimentation/Amazon
Amazon logo
Amazon
Aug 14, 2026
easyBusiness Intelligence EngineerTechnical ScreenAnalytics & Experimentation
0
0

A mobility service tracks trip duration, trip fee, and a post-trip survey outcome recorded as good or bad. Design a visual analysis that progresses from one metric to two, three, and four variables. Explain which charts you would use, what each reveals, and how you would avoid misleading encodings or overplotting.

For the fourth variable, choose one useful dimension such as city, vehicle type, or time period and state why it belongs in the analysis.

Constraints & Assumptions

  • Duration and fee are positive continuous variables with possible right skew and outliers.
  • Survey outcome is binary and may be missing nonrandomly.
  • The audience needs both a high-level summary and the ability to diagnose segments.
  • Do not encode several continuous variables only by bubble area without explaining perceptual limitations.

Clarifying Questions to Ask Guidance

  • Is the goal operational monitoring, causal diagnosis, or executive reporting?
  • What is the observation grain: trip, rider, route, or daily aggregate?
  • Are fee and duration comparable across cities and vehicle types?
  • What fraction of trips receive a survey response?

What a Strong Answer Covers Guidance

  • Appropriate single-metric summaries rather than a context-free large number.
  • A scatterplot or density view for duration and fee with skew and overplotting controls.
  • A clear encoding for binary survey outcome and a comparison of conditional rates.
  • A justified fourth dimension through facets, small multiples, or stratification.
  • Missing-response bias, aggregation grain, and accessible color choices.

Follow-up Questions Guidance

  • How would you separate the effect of trip duration from city mix?
  • What chart works when there are millions of trips?
  • How would you communicate survey nonresponse to an executive audience?
Loading comments...