Time Serious

Time Serious

Organisation: Guardian Australia (Australia)

Publication Date: 03/04/2016

Applicant(s)

Description

First the problem: Look at Paul Farrell. He's a journalist. Sometimes, editors want Paul to do stories that are formulaic, such as covering stories about the weather - how hot is it, is it a new record, or recurring stories like politician's expenses - who spent the most, who are the top 5, etc. Paul wants more time to work on deeper, more meaningful stories - like why the climate is changing, which politicians are rorting the expenses system, or why the AFP is following him around. Now, the solution: Many of these stories have a common element - they are based on periodically released time series data. We surveyed the last week of stories in the australian news section of the Guardian, and found at least five such examples alone. Economic figures, polling figures, emissions data, and weather data. In a single sentence, we want to make a system that takes a data input, analyses it, and outputs a news report in natural language, with publication-ready graphics. While similar things have been built before, as far as we know these are either proprietary and closed, or they've been built for specific data input, like earthquakes or sport stats. Ours will be open source and data agnostic. The outcome: You might be worried that we'll make Paul redundant. In fact, he now has more time, and can work on deeper, more meaningful journalism, or can add context to the stories produced by our system. The editor is happy as they're still getting their fast, top-line stories. Workflow: -Data is added to the system. This could be automated, such as by plugging the output from a scraper -The program analyses the data, covering the typical building blocks of these news stories: some examples are the largest and smallest values, averages, largest ever values, outliers, percentage increases year on year, and so on. -The results are outputted in two ways - one, generating a news-style explanation of the data. We will also generate graphics for the dataset, styled and ready for publication

Technologies used for this project:

Python, Flask, Javascript, Google Docs
Follow this project
Wait
Comment

Comments (0)

You have to be connected to contribute

You have to be connected to follow

Leave this project and no longer be informed about this project

By joining this project, you will be informed by email when an update or a new contribution is posted on the website.

Thank you for your active participation !
Best,

The GEN Community Team