Insights for Action

Data Sources and Cycling

Thursday, 3 Sep, 2020 By Paul Hewson.

Something I’ve been droning on about for a few years concerns our reluctance to use datasets that already exist.

Specifically in relation to cycling, we have a lot of data on cycling that is rarely used, and that’s without worrying about data that aren’t disseminated in a way that lets us use it. Here are some high quality UK based data sources. These come from surveys that are very expensive to conduct:

  • Travel to work data (Census, 2011).
  • National …

Privacy by Design

Wednesday, 2 Sep, 2020 By Paul Hewson.

Data Controller image Photo Credit: Photo by ThisIsEngineering from Pexels

Legislation which carries large fines usually provides a good reason for checking out the law. In this case, the driver is the General Data Protection Regulation (GDPR), an EU directive which addresses the fundamental rights and freedoms of natural persons. Natural persons are “real” people, unlike llegal persons which includes entities like companies. It asserts that natural persons have the right to …

Reported Road Injuries in GB

Friday, 7 Aug, 2020 By Paul Hewson.

In the event of a road collision involving at least one vehicle and one injury, there is a requirement in the UK to report to the police within 30 days. The data are called “STATs19” after the form on which the data are collected, and are made available as “Open Data” for anyone to analyse.

This immediately raises a number of points:

  1. What is a “road” collision (do car parks, driveways, private roads count)?
  2. What is an “injury”? More specifically, injury severity is coded …

Insights from Data

Thursday, 6 Aug, 2020 By Paul Hewson.

Lightbulb image Photo credit: Vojtech Okenka from Pexels

Having worked as an academic, as a Data Scientist, as a Business Intelligence Specialist as well as doing a lot of consulting I have learnt a lot about the way we use data, and the way people perceive “statistics” (whatever that word really means).

Sometimes, a problem is so straightforward it only requires technical input (data processing) and it’s solved. More often though, a lot more interpretation is needed. …

Mathematical models and statistical models

Thursday, 6 Aug, 2020 By Paul Hewson.

I suppose I started thinking about this really hard in the early days of the COVID-19 pandemic. I’m not sure it’s something we articulate well; the difference between a mathematical model. A mathematical model is meant to capture the science of a phenomenon in mathematical terms. A statistical model is meant to “fit” data and account for the uncertainty; but doesn’t necessarily explain the science. And there are certainly plenty of examples where you’re fitting a …