1.  Wikidata is valuable as a model for how to organize properties on your own Wikibase install, and finding ideas for how to use your data.  It is also useful as an easy to extract data from source to start building your own Wikibase install.

(II) Please indicate 1 thing that you particularly like from Wikidata, and 1 thing that you don’t feel so comfortable with from Wikidata

2.  Wikidata is a lot like old school library catalogs.  Information is near each other and can suggest alternative pathways or relationships that can be worth exploring when writing academic papers.  

One thing I do not like are generic nature of some properties which make it hard to use domain specific knowledge in a way that is easily extractable for queries or knowing what should be entered.  An example is classification.  For Paralympic sports, this is the type of disability classification used in competition, but it can also refer to classes of submarines.  For education, it can mean university or a non-educational sports organization where someone got a scholarship to sponsor elite sport.  This can require huge domain subject knowledge to clean up.

Things I do not like are unsourced statements, which make it hard to judge value of data as many numbers have a subjective value based on who created them.  For example, how many people died as a result of World War I.  Better to have multiple numbers with multiple sources than one source, even if they disagree.  Death toll from Hurrican Maria is another example where multiple numbers is a good thing, even if there is source conflict.

what else can Wikidata do to keep / improve data quality

3.  Quality can be improved by adding more references to statements, and realizing that one statement on a property may not be enough.  Encourage people to source, especially when manually adding data.  When doing bulk data imports, make sure there is a quality aspect to it and seek out these sources.  Don't necessarily wait for people to offer data. 

how do you think we can get more newcomers involved?

4.  Get more newcomers involved by thinking more about data output.  Wikidata items are meant to be read by machines, not by humans.  Humans need human readable output.  That output needs to be created in such a way that it is useful.  The query engine creates to many barriers to entry for data use.

Also, realize that different users have different needs.  Think about needs of specific user groups and how to address the needs of that group.  Wikipedians, academics, edit-a-thon organizers and developers all have different needs.  By addressing those needs more specifically, you can better reach these audiences.

Sincerely,

Laura Hale