Project avatar for Living with Machines

Living with Machines

Finished! Looks like this project is out of data at the moment!

See Results

Thank you everyone who helped complete our projects!

Results

Article

See Talk post from 2024.

Results from our workflows were included in our exhibition, co-curated by the British Library and Leeds Museums and Galleries, open at Leeds City Museum from July 2022 to January 2023:

Follow our progress

We'll post regularly to our blog and tweet from @LivingWMachines with news about the team and our work. We'll also tweet when we add new images or tasks to the site. We're also posting on Mastodon at @LivingWithMachines@zirk.us

Sign up to our newsletter for occasional updates direct to your inbox.

Improving workflows

We have set up the project so that the results of your classification tasks directly feed into annotation tasks. For example, images you mark as unreadable are dropped from the next task, and articles classified as describing industrial accidents feed into specialist annotation tasks that create data we can use in our analysis. This also means we're only asking you to annotate relevant articles that have passed through the classification workflow.

Sharing results

Our results include methods, datasets and research findings. We're publishing our results as peer-reviewed papers, blog posts, presentations, software models and tutorials on the British Library's Research Repository and elsewhere.

All contributions to this project by Zooniverse volunteers are archived permanently in the British Library's research repository and made available for reuse - by anyone and for any purpose - under an open access public domain license. The first two datasets created by Zooniverse volunteers are available on the British Library's research repository.

Using the results

As an experimental project with research questions and goals that evolve over time, some uses are likely to change as we develop our methods and learn from our results. Our initial uses include:

  • Improving the tools we are developing to find industrial accidents in newspapers by helping us understand how to combine keywords searches with computational linguistic processes.

  • Articles classified by participants may be used as a 'ground truth' dataset to train machine learning software to find other articles about industrial accidents in the wider corpus of digitised newspaper articles.

  • Analysing the language in articles to understand how it changed over time, topic and location.

  • Transcribed names may be used to improve automatic text transcription tools (particularly optical character recognition, or OCR).