Manually code sample of data

OpenDataLiteracy / Feedback

Research project on feedback, question and answer, and user comments on open data portals

0 stars 1 forks source link

Manually code sample of data #4

Closed nniiicc closed 7 years ago

nniiicc commented 7 years ago

I will send this by email and we will review the classification April 21, 2017.

nniiicc commented 7 years ago

Sample issue for coding here: https://docs.google.com/spreadsheets/d/1lugxssdgS_LaspCsA8JAMqiPuI13QwoN30a7BxAac3g/edit#gid=1598819504

annieyan commented 7 years ago

Some observations/thoughts from the literature:

the possibility of one comment belongs to multiple issues
the issues have a hierarchical structure maybe
user generated tags might be dirty and random. Predefined categories might be better
if possible, we could manually label all of them

annieyan commented 7 years ago

Maybe categorize the issues to broad categories such as "data availablility""data quality""metadata quality". Then we use finer grained sub-categories in our classification. For example, broken link might belongs to data availability...look forward to your thoughts on this