[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
Page MenuHomePhabricator

score/scoring explanation
Closed, ResolvedPublic

Assigned To
None
Authored By
Lydia_Pintscher
Feb 28 2021, 3:02 PM
Referenced Files
F34196152: image.png
Mar 29 2021, 12:32 PM
F34131660: image.png
Mar 2 2021, 2:10 PM
F34131657: image.png
Mar 2 2021, 2:10 PM
F34127888: image.png
Mar 1 2021, 12:21 PM

Description

As an editor I want to get a quick explanation of the score and the scoring in order to be able to understand the result.

Problem:
We should not expect the user of the tool to understand what the score means. Additionally we need to make it clear that we are judging the quality of the Item and not the quality of the entity described in the Item. This is especially important for Items about people. We are judging the data about the person but not the person.
We need to provide a short explanation for this.

Screenshots/mockups:
TODO

BDD
GIVEN
AND
WHEN
AND
THEN
AND

Acceptance criteria:

  • quick intro text exists for what the score means and what goes into it

Event Timeline

Currently we have

image.png (60×398 px, 8 KB)

If this is not sufficient, I can suggest additional text. However, I assume that explaining this will take more than just one line, so it will withdraw focus and attention from other relevant elements in the UI.

Currently we have

image.png (60×398 px, 8 KB)

If this is not sufficient, I can suggest additional text. However, I assume that explaining this will take more than just one line, so it will withdraw focus and attention from other relevant elements in the UI.

Yeah I think we need 1 or 2 sentences about what factors into the score at least and what doesn't influence the score.
Maybe we can make the learn more link not lead to the ORES page but instead open a popup or flyout or similar?

We could have a "How is this score calculated" flyout and put all info about ORES, difference between the things items refer to and the content of the items themselves etc. inside there.

I considered the flyout, but it turns out to be harder to integrate in existing layouts, so I went ahead with a popup.

image.png (491×565 px, 88 KB)

It would be called from the link in the header on both pages.
The page is not scrollable when it is open; the popups position is fixed. Thus, the popup has a scrollbar.
It closes again when clicking on the "Back to…" on top, or, on desktop, when users click on the darkened areas around it.

The popup is opened from a link in the header on both of our views

image.png (65×554 px, 12 KB)

@Jan_Dittrich I was wondering why the explanation popup has a link "back to..." instead of a "close" and/or "x"? thx

There is no x to keep it easy to use (aka hit with your fingers and find it at an expectable position) on mobile. I could try to find out what Wikit is doing with popups and if they change between mobile and desktop and suggest a matching behavior here.

There is no x to keep it easy to use (aka hit with your fingers and find it at an expectable position) on mobile. I could try to find out what Wikit is doing with popups and if they change between mobile and desktop and suggest a matching behavior here.

Ah I see, thanks for clearing that up. Yes that would be great if the behavior matched.

Here is the text from the mockup above:

The Item Quality Evaluator retrieves a quality score for each requested item and also calculates the avarage score of all items requested.

*What does the score measure?*
The score estimates the quality of the Wikidata item. It can not estimate if the content should be included on Wikidata or if what the item is about is “good”: The item “famine” (Q168247) can have a high score if it has references and delivers complete information, while “Kitten” (Q147) can score low if it lacks references and statements.

*How is the score calculated*
We trained the ORES Algorithm by giving it items that we scored to belong in one of 5 quality categories. Now, if the algorithm is given an item, it estimates the the likelyhood that a given item falls in each of the 5 possible quality categories. To get one score for an item, we weight each category by its likelyhood and calculate an average of the weighted scores, which is the score for a single item.

From several single-item-scores, we calculate the avarage score of all the items that you requested.

However,

*Where can I find more information?*
This tool is documented at [somewikilink]. Its source code is on [platform].
If you want to know more about [ORES] read [wikilink]

Obviously needs still the right links

So, as I learned from a team member the WMF has put up their modal design. They use a right-upper corner x.

image.png (1×1 px, 290 KB)

This is the implementation from Bootstrap, X on the right with a title.

I have changed the wording for the content under "Where can I find more information?" to:

This tool is documented and the source code is available on GitHub[link]. (as they share the same link)

You can read more about ORES[link] in Wikidata in this blog post[link]. (I got this from https://www.wikidata.org/wiki/Wikidata:ORES)

Botoxparty moved this task from In review to Done on the Item Quality Evaluator board.