ChatGPT | Online Journalism Blog

Robot with books — Image by kjpargeter on Freepik

I’m speaking at the Broadcast Journalism Teaching Council‘s summer conference this week about artificial intelligence — specifically generative AI. It’s a deceptively huge area that presents journalism educators with a lot to adapt to in their teaching, so I decided to put those in order of priority.

Each of these priorities could form the basis for part of a class, or a whole module – and you may have a different ranking. But at least you know which one to do first…

Priority 1: Understand how generative AI works

The first challenge in teaching about generative AI is that most people misunderstand what it actually is — so the first priority is to tackle those misunderstandings.

Continue reading →

If you’re working with data as a journalist it won’t be long before you come across the phrases “dirty data” or “cleaning data“. The phrases cover a wide range of problems, and a variety of techniques for tackling them, so in this post I’m going to break down exactly what it is that makes data “dirty”, and the different cleaning strategies that a journalist might adopt in tackling them.

Four categories of dirty data problem

Look around for definitions of dirty data and the same three words will crop up: inaccurate, incomplete, or inconsistent.

Dirty data problems:
Inaccurate: Data stored as wrong type; Misentered data; Duplicate data; abbreviation and symbols.
Incomplete: Uncategorised; missing data.
Inconsistent: Inconsistency in naming of entities; mixed data
Incompatible data: Wrong shape;
‘Dirty’ characters (e.g. unescaped HTML)

Inaccurate data includes duplicate or misentered information, or data which is stored as the wrong data type.

Incomplete data might only cover particular periods of time, specific areas, or categories — or be lacking categorisation entirely.

Inconsistent data might name the same entities in different ways or mix different types of data together.

To those three common terms I would also add a fourth: data that is simply incompatible with the questions or visualisation that we want to perform with it. One of the most common cleaning tasks in data journalism, for example, is ‘reshaping‘ data from long to wide, or vice versa, so that we can aggregate or filter along particular dimensions. (More on this later).

Continue reading →