mass edition: clustering, complicated split and detach

[paulgirard]

Three data edition features need potentially to leverage a mass edition User Interface.

Mass edition means:

• long list of items to treat
• filters to squeeze that list
• infinite scroll to go through the list
• advanced selection feature (select all, selection cart...)
• actions to be applied on a selection

The data edition features which need this:

Clustering

As a researcher I need to review part of the dataset to make sure duplicated names of people, organization and streets are edited in order to rely on accurate numbers for my research.

Scenario:

• a filterable list of people with filters on a side
• people are grouped by alike names forming groups of potential duplicates
• I need to review group composition and take decision about them
• I need to :
  • exclude some member of the group
  • review some people page to learn more about them
  • merge the group by choosing a common name for all of them (existing merge feature)

Question : does the merge apply only on messages corresponding to the filters or on all?
Only on filters

Complicated split

As a researcher I might need to be able to split one item into many to disambiguate one homonym into two different people. Yet those should never appear in the same message together as they are not active on the same period of time in spite of the common name. Thus I need to chose a set of message to be attribute to the new people 1, leaving the rest of the messages to the new people 2.

Scenario:

• a filterable list of messages with a filter active on the people to split
• I create people 1 and people 2 in the split destination form
• I add a time filter to reduce the messages list into the period of time when people 1 was active
• I select all (even if not viewable in the list as too numerous) and decide to attribute them to people 1
• I change time filter to meet period activity of people 2
• select all
• attribute to people 2

Question: how to make sure all message are attributed?

detach

Same as split ?

tags

Assign tag on large number of items