Course
Advanced Trove-ing

Draft, 03 March 2017
This page should be in a useful state, but still needs work before it's finished.

How well do you know your Australian history?

Headline Roulette

Headline Roulette presents a randomly selected newspaper article from Trove. It’s built using the Trove API.

So what is Trove?

Trove

  • A platform for delivering digital content from the NLA and partners – access to full text, images, and more.
  • A collection of collections aggregating records from around the country for easy discovery and use – access to metadata only (but you can often click through to the original record).

All this content (over half a billion items) can be discovered through Trove’s zones. Keep in mind that some resources appear in multiple zones.

Digital content

  • Newspapers (and Government Gazettes) – not just the big metropolitan papers, but many small local papers, community, political, and religious papers as well. How about Chinese language newspapers?
  • Browse titles by place
  • Trove digital library – books and journals. Searches will return individual articles from digitised journals – note that you’re actually searching the OCRd text of the book or article, not just the catalogue data! Try searching for “Dificit unaccounted” and following the links to see how your query is passed on to the ‘inside’ of the work.

Large amounts of high-quality digitised material have been added in recent years, such as deep-zoomable images and maps.

  • For other zones try typing nuc:"ANL" in the search box and checking ‘Available online’ (or clicking the ‘online’ facet)
  • For example, explore the photos and maps
  • If your searches for NLA digital content are returning a lot of noise (such as books in the Maps zone) try adding “nla.obj”.

A collection of collections

Trove includes collection records from:

  • libraries, museums, archives, galleries – small to large
  • government, research agencies, university repositories
  • community groups, and even individuals (via Flickr)

Some things you might not expect to find:

Some search tips

Search tips

For more detailed information on finding things in Trove, remember to look in the Trove Help Zone.

Start big, then narrow your results

There’s so much stuff in Trove that it’s tempting to think your search should be as precise as possible. But it’s often more effective to start with a broad search and then narrow things down using filters and facets.

  • There are two main filters – ‘Australian’ and ‘Online’. You can activate them using the checkboxes underneath the search box.
  • The ‘Australian’ filter attempts to focus on material either by Australians or about Australia.

  • The ‘Online’ filter focuses on material that is available online. There’s often a bit of uncertainty about this because the metadata Trove receives is not always accurate. You can limit further by online status using the facets.

  • Facets filter on characteristics of the metadata, so different zones will have different facets.
  • In the newspaper zone, you can use facets to limit your results to a particular state, or selected newspaper titles (and lots more).

“Exact” searching

Another way of focusing your search is to change the level of “exactness”. You probably already understand the difference between a search for keywords and a search for a particular phrase:

  • white australia – returns results including the words ‘white’ and ‘australia’.
  • "white australia" (note the double quotes) – returns results including the phrase “white australia”.

But you might notice that the phrase search also returns results that match “whiter Australia”. Within the phrase, words are ‘stemmed’ to find all forms of that word. You can stop this by preceding the quotes with text:.

In the newspapers zone there’s also a bit of extra wriggle room to deal with some of the vagaries of OCR. This means that some results might include an extra word in the middle of the phrase. To switch this off add ~0 after the quotes.

Here are some comparisons:

Query Number of results Explanation
white OR australia 40,855,944  
white australia 4,971,530 Same as white AND australia
"white australia" 136,327 Search for phrase (with stemming)
text:"white australia" 132,470 Search for phrase (no stemming)
"white australia"~0 130,164 Search for phrase (no extra words)
text:"white australia"~0 126,982 Search for phrase (no extra words and no stemming)

Search box super powers

Any of the ‘special’ searches above can just be typed into the search box. In fact, you can build up quite complex search queries just using the search box.

  • Boolean searches – use AND, OR, and NOT, to specify how your keywords or phrases should be combined. For example, try "white australia" AND chinese versus "white australia" NOT chinese.
  • Limit your search to the collections of a particular organisation by usin”g nuc: and the organisation identifier (in quotes”. For example nuc:"ANL" searches only the NLA, while nuc:"UC:IR" searches only the University of Canberra’s institutional repository.
  • Want to limit your search to things that have user tags? Try has:tags. Also works with comments and corrections (in the newspapers zone).
  • Proximity searches specify the maximum number of words you want to appear between your keywords. For example (cat dog)~5 will find articles where the words ‘cat’ and ‘dog’ are a maximum of 5 words apart.

Finding photos you are free to use

Organising and sharing your research

Trove tags

  • Download resources for offline use – images, PDFs, text
  • Tags – tag items to make them easier to find
  • Lists – save your discoveries in thematic lists
  • Use Zotero to save Trove resources to your own research database.

Trove as data

QueryPic

Say hello to the Trove API

Big pictures

Getting your data in bulk

Integrations

  • Europeana WWI – uses APIs to create an federated search across Europe, USA, Australia, and NZ.
  • Austlit – automatically retrieves newspaper articles related to publications, eg The Land of the Sun
  • Atlas of Living Australia – automatically retrieves resources relating to species, eg. Magpie
  • TungWah Times
  • Loom index – cool new subject browser from SLNSW’s DXLab
  • TroveNewsBot – search Trove from Twitter!

Creative projects

Even non-coders can make things with the Trove API

Tags

Trove