10 Rapid prototyping: Data analysis

 

This chapter covers

  • Prototyping ideas quickly to support a business case
  • Exploring a dataset to build a proof of concept

Sometimes, data analysis is about investigating whether an idea is viable. For example, is the available data good enough for the business to build a data-driven app? We can answer questions like this one by analyzing the data, but it would be even more powerful if we could build a working proof of concept. By doing this, we make the idea come to life for our stakeholders. Crucially, we also discover any obstacles to using the data for this purpose.

Real business case: Building a proof of concept

This chapter was inspired by one of the first projects I delivered as a data scientist. The data team was tasked with creating a revenue-generating product. The idea was an app that lets someone enter a vehicle’s registration and display the current market conditions for that vehicle and vehicles similar to it. Defining what made a vehicle similar enough was the hard part, and that was the data-driven secret sauce behind the product. Once we’d established some rules, I built a quick proof of concept to show stakeholders that the idea worked. Building this proof of concept also meant I could communicate the specific challenges of building the product for real, which I knew because I had actually tried to make it work with the available data and systems.

10.1 The rapid prototyping process

10.1.1 Rapid prototyping example

10.2 Project 7: Build a proof of concept to investigate Welsh property prices

10.2.1 Problem statement

10.2.2 Data dictionary

10.2.3 Desired outcomes

10.2.4 Required tools

10.3 Applying the results-driven method to investigating Welsh property data

10.4 An example solution: Building a prototype to explore using house price data

10.4.1 Analyzing data before prototyping

10.4.2 Investigating geographic aspects of a dataset

10.4.3 Identifying how to present data in the prototype

10.4.4 Project progress so far

Summary