User:Multichill/No claim

This page describes the approach I'm taking to get rid of no claim items.

Introduction

Every item on Wikidata should have at least one claim linking it to another item. Without claims we don't anything about the item. The number of items without claims peaked at almost 9 million items ([1]) and seems to be slowly decreasing.

To further reduce this I try to add at least one claim to items based on the templates and categories used on a connected Wikipedia article. I first tried this on the Dutch Wikipedia and later did the same on the English Wikipedia

Some statistics

I ran some statistics to figure out what to work on:

site	Items
enwiki	562642
jawiki	227952
frwiki	177748
dewiki	173342
zhwiki	154622
eswiki	143011
plwiki	139857
itwiki	137467
ruwiki	117607
ptwiki	111771

This is base on this query which produces a long list.

Make the list

The first step is to do a query to find all items that have no claim and that do link to the Wikipedia we're working on. I'm using Dutch as the example because the English lists are still huge. The query and the result.

Put the list on the Wiki

I put the list on the Wiki so that the links are in the database. Later on we'll use that for queries. The list is split by starting letter. This seems to make the (27) lists small enough to be able to save the pages. You can see this in Dutch and English. I can do this by hand, but also have a small bot to automatically do this.

Find templates to claim

Each template on a page will give you a big of information of what the page is about. First I do a query to figure out the most used infobox templates (Dutch and English example). Based on that I make a page with tells a bot per template what claim(s) to add. This contains of a wiki per with per line the name of the template and the P/Q combination to add (Dutch and English example). Finally I do a query to create the full list and run a bot to add the missing claims.