Documenting Your Data
Some tips that will make your life easier.
Felix Rau University of Cologne
Why record metadata?
It is important, just ask the NSA.
Metadata procrastination
“I will do that later.”
“Ah, I don’t need to write THAT down.”
“I will remember that when I’m back in my office.”
You don’t have to love it
... but please do it.
Worst case scenario
- You need information to understand data
- Other people don’t know what you know
- Future-you doesn’t know what you know
- To find the right bit of data you need information about your data
Metadata – Data about data
Types of metadata
File characteristics
Descriptive Metadata
Structural Metadata
Administrative (incl. rights)
Make your life easier
- Record metadata early
- Create data for bulks of files
- Document raw data and processed (transcribed) data
Deal with your raw data
File related Metadata
- File Type: WAV, MPEG, ...
- Rename your files (Include date, language, etc. in the name)
- Create metadata for bulks of files
Make an inventory
e.g. with exiftool
- File name
- File type (e.g. WAV)
- File size
- Duration
- Encoding, sample rate, bit rate
- Dates of creation/modification ...
Metadata for your processed data
Metadata formats
- IMDI – used by the LAT-Archives such as Nijmegen, Cologne, and Mexico D.F. (CIESAS)
- CMDI – CMDI is a modular format and IMDI-CMDI and ELDP-CMDI
- Forms or spreadsheet of different archives AILLA
- ...
IMDI/CMDI
- Resources (files):
audio and video files, annotations, ... - Sessions (set of files):e.g. an audio file with ELAN file and a toolbox file
- Corpus (set of sessions):
sessions joined as group
CMDI Maker
CMDI Maker
Corpus
Resources
Create Sessions
Sessions
Actors
Sessions
Export
Further editing
CMDI Maker
an offline enabled web app
- works in modern browsers (best in Firefox, Chrome)
- no installation, just go to http://cmdi-maker.uni-koeln.de/
- data is stored persistently (even if you close the browser)
- works offline
En Español
Internationalisation
More information:
In case you need more help:
Email the team: cmdi-maker-support@uni-koeln.de
... or just email me: f.rau@uni-koeln.de
Summary
- Create your metadata early
- Use some documented format
- Use the available tools