You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
A few ramblings about discoverability.... to eventually work into the guide. People tend to not read long lists of things. Once collections of datasets are > 30-60 items, you need the ability for the user to filter. Both the website owner & dataset supplier have a role to play here. Metadata tags are commonly used. They can be list of keywords, key value pairs, mandatory or optional. Human-provided or natural language processing generated from text describing the dataset. In both the dataset supplier should provide more rather than less. Website user should consider whether public can suggest or add tags as needs will change orerely be beyond what dataset supplier originally envisions. One frame of reference to consider is when there are 5000 different datasets how do you quickly get the most useful datasets to appear for users of different skill levels, who have different time-to-start thresholds, different machine-learning interests, different geoscience interests, and different sample size needs.
The text was updated successfully, but these errors were encountered:
A few ramblings about discoverability.... to eventually work into the guide. People tend to not read long lists of things. Once collections of datasets are > 30-60 items, you need the ability for the user to filter. Both the website owner & dataset supplier have a role to play here. Metadata tags are commonly used. They can be list of keywords, key value pairs, mandatory or optional. Human-provided or natural language processing generated from text describing the dataset. In both the dataset supplier should provide more rather than less. Website user should consider whether public can suggest or add tags as needs will change orerely be beyond what dataset supplier originally envisions. One frame of reference to consider is when there are 5000 different datasets how do you quickly get the most useful datasets to appear for users of different skill levels, who have different time-to-start thresholds, different machine-learning interests, different geoscience interests, and different sample size needs.
The text was updated successfully, but these errors were encountered: