Needs Assessment

[drjwbaker]

Build on the July workshops with a short needs assessment, perhaps emailed to previous attendees.

Potential (short) exercises:

• rank our ideas for practical use of CL methods (structured data)
• ask (having attended) what tasks people would like to do that they think CL could help with.

Actions:

• Draft poll
• Decide platform for polling.
• Design poll
• Contact attendees
• Review results

[drjwbaker]

@rossi-uk My first stab at this below (as promised). Comments/suggestions welcome!

---

Needs Assessment for computational analysis of catalogue data

Having attended the training session 'Computational Analysis of Catalogue Data', we now need your help refining our focus as we develop the training materials over the coming months.

This short needs assessment focuses on the practical applications of computational analysis of catalogue data.

It should take no more than 5 minutes to complete.

Q1. Has the training session helped you achieve anything specific?

Q2. (multiple choice) Our training materials will conclude with two or three optional 'next steps' modules (15-20 minutes in length), each of which is intended to develop you ability to undertake complex analyses of your catalogue data, in most cases beyond AntConc. Based on what you know, please rank the following on the scale of most useful (1) to least useful (5)

• Named Entity Recognition.
• Analysing Word/Lexical Choice.
• Comparing Catalogues
• Comparing Catalogues to standard language corpora (e.g. for British English, the British National Corpus)
• Using regular expressions in AntConc

Q3. Please briefly describe a catalogue related task that you think could be improved or made possible by the use computational analysis/linguistics.

Q4. Any other comments.

[drjwbaker]

@rossi-uk I've drafted this at https://docs.google.com/forms/d/1nZgEyMR0htXc4SkzwGlu0vK0RWnU_AEEOeqH09s6qqc/edit If/when you are happy, I'll send out to attendees.

[drjwbaker]

Circulated with polite request to complete by 25 September.

[drjwbaker]

Results

6 responses. Data on drive

Q1: Has the training session helped you achieve anything specific?

50/50. Which feels a good outcome.

Q2: If you answered 'yes' to Question 1, please briefly describe what that might be.

Identifying phrases/words. Emphasis on insensitive language.

Q3: Our training materials will conclude with two or three optional 'next steps' modules (15-20 minutes in length), each of which is intended to develop you ability to undertake complex analyses of your catalogue data, in most cases beyond AntConc. Based on what you know, please rank the following on the scale of most useful (1) to least useful (5).

Named Entity Recognition the clear winner. Regex the clear loser. Comparative analysis the next best.

Q4: Please briefly describe a catalogue related task that you think could be improved or made possible by the use computational analysis/linguistics.

Lots of rich detail here but a few key points:

• variants of spelling/words
• flagging words/records for attention/revision
• named entities

[drjwbaker]

This is done. Note that feedback will be useful in end of project reporting.