jeffsteward / mcn2017

Many institutions have been releasing their collections on GitHub and sharing APIs over the last few years.

I'd like to propose a hackathon format where museum technologists can get their hands on real-life datasets from other museums and share their experience dealing with open data.

I recently had some spare time in between projects and have taken a deep interest in text-mining and data-mining using Java, Python and R. I would be excited to take part in something about this during MCN.

The ideal would be to involve a local museum.

#OpenAccess, now what? -- half-day data hackathon proposal for #MCN2017

This year MCN will focus on how museums can use technology to innovate and emphasize transparency, individual action, and institutional bravery. (from the MCN2017 call for proposals)

Title: #OpenAccess, now what?

Get your hands on datasets, text-mining, data-mining, dataviz tools and tricks. A #MCN50/#MCN2017 data hackathon for museum technologists @MuseumCN.

Format

• a half-day hackathon before the conference (or should it be a full day?)
• a follow-up session during the conference to present the results

in others category

What?

This is a hands-on data crunching session.

This isn't a workshop (there isn't one workshop leader, each participant is bringing and learning something) but it rather takes the shape of a hackathon where all participants share tips and techniques and produce a few stunning presentations by the end of the day.

Some results could be presented in a session later during the conference.

Take-aways for "datathon" participants

• Evaluation => What can your data tell you about your visitors and your collection? How can this feed your digital strategy?

• Data-led storytelling => Can data help you decide where to focus your interpretation efforts?

• Data-backed storytelling => How can data not lead but strengthen and support your narratives?

• Accessibility => How can you increase accessibility with big data and machine learning? (eg. reducing costs of translations)

• Strategy/etc => What do you get from "opening" your collection? Feedback from institutions that have just done it.

One of the principles of data science is that it has to be reproducible. So, one outcome could be to share a repository of tools on GitHub. A data toolkit that can be applied to "any" museum dataset.

When?

Either:

• a half-day or a full day on workshop day, before the conference officially starts
• or the week-end before MCN50 in Pittsburgh, if a local museum gets involved

How?

• a series of short lightning talks where participants present tools or techniques they use, and how they use them, to feed the discussion
• participants/groups choose to apply some of these tools and techniques to a dataset

Who?

2 profiles:

• people at ease with programming and hacking code, be it with Java, Python, R, and/or dataviz techniques

• people who can bring a large dataset and are curious to explore it in a different way (eg. a collection of publications, labels, audio guide scripts etc.)

Lightning talks and expression of interest to participate in this hackathon

• Seema Rao : on setting goals when working with data
• Christophe Buffet : on using Java for text-mining
• Jeff Steward : on Harvard Art Museums’ statistical landscapes (see Suns Explorer, Collection Activity, Collection Terrains)
• Chad Weinard : on releasing The Williams College Museum of Art collection data (see New Dimensions for Collections at WCMA)
• Jeffrey Inscho and the CMP_studio space/team

People to involve

• MCN’s Data and Insights SIG
• many more people from the MCN community and outsiders

Structured data sources (datasets available)

• CMOA
• WCMA
• The Met
• Cooper-Hewitt
• MoMA
• MoMA's exhibition and staff histories
• Tate
• ACMI
• ACMI Historic Film Screening Data
• others can be added to this list, cf. gitMuseum

Museums Collections APIs

• API Harvard Art Museums
• API Collection of Cooper Hewitt, Smithsonian Design Museum
• Brooklyn Museum Opencollection API
• Collection API - Science Museum Group
• Rijksmuseum API
• Walters Art Museum Collections API
• and many more
• CHECK if there is a collection API for the Barnes? Barnes Foundation

Unstructured data sources (meta topics)

• MCN-L archive
• twitter hashtags such as #musetech, #mtogo, #MCN2017
• twitter handles such as @MuseumCN
• papers eg. MW papers
• conference programs
• conference participants lists
• conference presentations slides or transcriptions
• websites eg. MCN
• emails
• etc.

Dataviz examples

• Dear Data
• FiveThirtyEight's view on The Met's collection: An Excavation Of One Of The World’s Greatest Art Collections by Oliver Roeder
• FiveThirtyEight's view on MoMA's collection: A Nerd’s Guide To The 2,229 Paintings At MoMA by Oliver Roeder
• MoMA Exhibition Spelunker by Good, Form & Spectacle
• about Jeff Steward’s “Suns Explorer” for Harvard Art Museums: When big data meets art appreciation
• Data Visualization Workshop Series at NCSU Libraries

Further reading

• Merete Sanderhoff — Open access can never be bad news
• Merete Sanderhoff — Your imagination is the only limit
• George Oates — on the MoMA Exhibition Spelunker
• Giorgia Lupi — Dear Data has been acquired by MoMA, but this isn’t what we are most excited about.
• Elina Sairanen — What Does Data Have to Do with It?
• Keir Winesmith and Anna Carey — Why Build an API for a Museum Collection?
• Chad Weinard — New Dimensions for Collections at WCMA