DATA IN THE DIGITAL HUMANITIES Michael Pidd 26 th November 2014, ICOSS, University of Sheffield. NatCen Seminar Series on Methodological Challenges

DATA IN THE DIGITAL HUMANITIESDATA IN THE DIGITAL HUMANITIES

Michael Pidd

26th November 2014, ICOSS, University of Sheffield. NatCen Seminar Series on Methodological Challenges

http://methodologicalchallenges.group.shef.ac.uk/

The data lifecycle in a typical digital humanities project:

1.Acquisition (e.g. digitisation)2.Processing (adding value)3.Analysis (and dissemination)

Data in the humanities is usually:

1. Small (discrete sources created by individuals).

2. Broad (many different types of sources have to be assembled).

3. Complex (because humans are not spreadsheets).

Rarely ‘Big’.

http://hridigital.shef.ac.uk@hridigital

Data acquisition:

1. Most of the evidence base is pre-digital. Very little is ‘born digital’.

2. Data acquisition is a question of translation, representation and interpretation.

3. The methods we use either enable or inhibit research.

4. But, the process also develops intimate knowledge of the evidence.

British Library NewspapersKeyword search for “pidd” gives 2,730 results…

Data processing:

1. Metadata can be complex, reflecting the complexity of the data.

2. Metadata can be very specialised, limiting re-use.

3. When processed at scale, computational methods are a trade-off between through-put and accuracy.

• Nominal record linkage using computational means to trace the lives of 90,000 people.• Record linkage across 45 separate datasets (some public, some commercial, all in different

formats and with different data models).• And most people have common names.

http://www.digitalpanopticon.org

Analysing data:

Do data visualisations tell us anything that we do not already know?

Data visualisation is only as good as the data.

Data visualisation should reveal trends and anomalies, directing us to deeper readings of the evidence.