Design an Original Trip-Data Analysis

# Design an Original Trip-Data Analysis Formulate one original question suggested by the green-taxi trip records and specify a reproducible analysis that answers it. Choose a question and methods that let you highlight your special skills and analytical philosophy, and explain that connection explicitly. In a live take-home, run the analysis and report the measured finding. In this self-contained practice version, define the exact result table or chart and explain how each possible observed result would answer the question without inventing unavailable measurements. ### Constraints & Assumptions - The question should not merely repeat the required distance, airport, tip, speed, borough-flow, or anomaly tasks. - State the population, outcome, time period, unit of analysis, and data-quality filters. - Treat exploratory findings as hypotheses unless a separate identification strategy supports a causal claim. - Make the technical or analytical strengths demonstrated by the chosen question visible in the handoff rather than adding complexity for its own sake. ### Clarifying Questions to Ask - Which available fields have been quality-checked? - Who could use the result, and which decision might it inform? - What result would contradict the initial hypothesis? - Which part of the question or method best demonstrates your particular skills and way of reasoning? ```hint Define the evidence first Write the rows, columns, comparison, and uncertainty display that would settle the question before exploring the data. ``` ### What a Strong Answer Covers - One precise and nontrivial question grounded in the trip fields. - Auditable eligibility rules, metrics, comparison groups, and uncertainty. - An exact result artifact and a conclusion rule tied to observed values. - A clear explanation of how the question and methods showcase the candidate's special skills and analytical philosophy. - Limitations, alternative explanations, multiplicity, and a useful next step. ### Follow-up Questions 1. How would you avoid selecting only the most interesting result after many exploratory comparisons? 2. What additional data would most reduce uncertainty in the interpretation?