2 From question to deliverable

 

This chapter covers

  • Preparing an end-to-end analytics project
  • Setting expectations with stakeholders
  • Managing the interpretation of results
  • Creating resources for reproducibility

Analytics projects begin with a question: How well is the business performing? Is the product easy for users to learn how to use? Did the marketing campaign produce a return on our investment? Questions such as these guide data analysis, statistical methods, and visualizations for communicating insights. The answer you provide to the question will ideally provide strategic information and direction to your stakeholders and their work.

While some analysts’ work is guided by their own questions, most provide insights into the questions posed by their stakeholders. These are delivered in a variety of ways—quick summaries can enable decisions, whereas larger projects may require statistical analyses and the presentation of results.

2.1 The lifecycle of an analytics project

With a question in hand, your responsibility as an analyst is to distill the organizational process, the research idea, or your curiosity into something you can define, measure, and report on. While some routine questions and analyses have well-structured metrics and data sources, most novel questions your team addresses will not have an available data source, metric, or statistical analysis method to guide your approach.

2.1.1 Questions and hypotheses

2.1.2 Data sources

2.1.3 Measures and methods of analysis

2.1.4 Interpreting results

2.1.5 Exercises

2.2 Communicating with stakeholders

2.2.1 Guiding the interpretation of results

2.2.2 Results that don’t support hypotheses

2.2.3 Exercises

2.3 Reproducibility

2.3.1 What reproducibility is

2.3.2 Documenting work

2.3.3 Exercises

Summary