Live notes, so an incomplete, partial record of what actually happened.
Tags: DH2018
Site: https://dh2018.adho.org/en/
My asides in {}
Tweet embeds are things I liked that seemed relevant to include or that captured things I missed
Building International Bridges Through Digital Scholarship: The Trans-Atlantic Platform Digging Into Data Challenge Experience / Construyendo lazos internacionales a través de becas digitales: La plataforma Trans – Atlántica descubriendo experiencias en desafió de datos
Elizabeth Tran, Crystal Sissons, Nicolas Parker, Mika Oehling
Trans-Atlantic Platform Social Sciences and Humanities Large group of funders in the partnership: strong in Central/South America and Europe (no longer includes direct EU funding) .. uses national funding rules .. T-AP Digging into Data was the 4th round of DiD, but first under T-AP .. consortia with 3 research groups across Atlantic .. currently preparing joint call on social innovation .. 109 proposals
Digging into Data Originated from US National Endowment of Humanities .. ideas that lots of good scholarship in the areas was outside the US or international .. challenge rather than just a grant: e.g., to answer the question "what you do with a million books / pages of newspapers / artworks?" .. but also foster interdisciplinary collaboration .. DiD is now over 10 years old. What have we learnt? Change in composition of countries/funder. Changes to international order making collaboration of this kind more important. Change in techL cutting edge stuff in 2009 now commonplace .. Round 1 (US, Canada, UK): inundated. 87 applications, 8 awards. Rounds 2 (Netherlands joined): 67 applications, 14 awards. Round 3: 69 applications, 14 awards. Round 4 (11 countries, 16 funders): radical departure from previous rounds due to sponsorship from T-AP - aims to change how funding worked. 101 applications, 14 awards. Move to 3 countries and across atlantic about ensuring new partnerships .. Challenges for Round 4: understanding commonalities and differences around peer review expectations, cultures of academic year calendars (e.g. long summers), different budget stability between funders .. Next steps: possible 5th round (NEH to offload organising if it goes ahead).
Joining fellow Trans-Atlantic Platform partners at #TAPWorkshopDiD4 on international cooperation in support to research projects in social sciences and the humanities. Mexican researchers from CONACYT-funded projects will be sharing best practices #DH2018 pic.twitter.com/eOzOuFw2Iw
— Arturo MenMar (@a_mendozam) June 25, 2018
Digging into Early Colonial Mexico
Came out of Spatial Humanities Lab at Lancaster and ambition to combine Geospatial Information Systems with textual processing: e.g. Natural Language Processing, Corpus Lingustics, Machine Learning .. Enabling mapping and spatial analysis of corpora .. 16th century historical documents .. data is the Relaciones Geográficas del Siglo XVI: very detailed records of Spanish colonies, people, environment, languages, et cetera.
#dh2018 Learning how the Digging into Early Colonial Mexico @DiggingCH project is getting along https://t.co/ble8rw1JAv Lovely project which encountered problems of NLP tools being optimised for English and poorly suited to indigenous languages. pic.twitter.com/0EKGSB2Fwa
— James Baker (@j_w_baker) June 25, 2018
Developing NLP/NER so that proper names for places in Old Spanish and indigenous languages can be found .. digital gazateer of C16 New Spain .. methodological questions are important as historical and geographical ones
Pre-1600 handwritten books across the world. Though mostly European. Provide international view of medieval manuscripts .. 3 bigish sets of manuscript data that don't play together at all .. Schoenberg Institute at Penn has the largest database of medieval manuscripts in the world .. Official repository for the Bodleian Libraries TEI-based Western Medieval manuscript catalogue. Great data, but only links out to one thing: VIAF. Paris and Penn have data on the same manuscripts, but different data based on different things that were considered locally important. For example, Oxford describes a manuscript on a shelf, Schoenberg describes ever observation of a manuscript (e.g. in a sales catalogue) .. most of the work thus far is on mapping across databases, and making local changes to start accommodating that mapping (e.g. using VIAF to populate fields) + lots of merging of place name soup .. Work resolving bad place name data has underscored to west of Poland bias of the data and the need to add new missing data.
Intelligent Search Engine for Belief Legends
Stories: murder, witches, ghosts. Folklore. Lots of these stories collected from late-C18 to early-C20. Dutch, Danish and German collections. Trying to rehabilitate the use of the map/space in the study of folklore. Developing network model to find communities in the data based on hypergraph search (nodes and hyperedges, latter represents relationships between nodes). Using OAI-PMH as a data model. Initial experiments with multi-lingual search didn't work. Instead doing dirty translated into each language using Google Neural Machine Translation, so search will bring back results based on those dirty translations + with domain specific keywords (12k list) thrown in. Doing TF/IDF with place names means you don't just end up mapping where people live (or in the case, where stories collected/told).
Machine Translation and Automated Analysis of Cuneiform Languages
Focus on Sumerian texts .. Cuneiform just a script used for writing in a number of languages .. small corpora presents challenges for machine translation ..
Escucho una presentación sobre para el procesamiento del sumerio y la escritura cuneiforme https://t.co/nTprmMFT0i #dh2018
— La Mónada é mobile (@epriani) June 25, 2018
.. Data from the Cuneiform Digital Library Initiative: info on ~500k artefacts, ~300k images, 2.65 million transliterated lines of text .. Ur III period: late third millenium BCE. Lots of bureaucracy therefore lots of texts .. annotation of corpora (taking into account disputes around translation) is a key part of the process.
Many work streams in ~30 strong team. Infrastructure: how do you do meaningful analysis across date digitised in different places for different reasons? Reprinting: how do information circulation within national corpora and beyond those corpora through translation? Movement of concepts: how did ideas grow and transfer? .. Mapping Metadata: conceptually similar but practically very different metadata across corpora. Create a map of metadata. Trying to come up with a technical ontology that someone else can use to understand how fields translate between corpora, and why the data is structured in the way it is. E.g. title: usually means different things in different places. Using OWL. Relied on institutional memory of digitisation and metadata creation .. Reprinting: case studies of events to track. Interested in geographic spread, networks of publications, and time dynamics. Building mini-corpora by hand at the start. Some big case studies: the eruption of Krakatoa (emotional history of news). Others more likely to only be part of specific international networks.
#dh2018 @epriani speaking about our @OceanicEx case studies of newspaper reprinting including volcanic eruptions and murder most foul. pic.twitter.com/puNM37CBTL
— M. H. Beⓐls (@mhbeals) June 25, 2018
Important to pick case studies that can be pinned down to dates - nothing on them before a certain date, very little after a certain date - in order to hand produce ground truth data .. Concepts: long time developments in the Atlantic world, relationship between abstract concepts and the specific languages in which they are expressed. E.g. Eugenics, using the infrastructure built by the project to understand how terms related to eugenics are used in debates. Role of diasporic media as node of information exchange.
Time for preliminary work with data is useful .. going to summer schools / conferences together as a method for creating shared knowledges/languages .. challenges in international collaboration around very different teaching loads, very different admin burdens of running a project .. granularity of responsibilities of project team important to get right .. having teams in place that then look for funding works well .. personal turnover a challenge: knowledge transfer needed each time ..
Things I'm learning from #DH2018 @DiggingIntoData projects workshop: data silos are still the norm, combining data reveals absences in coverage, research admin varies wildly by country, project plans are always optimistic, student labour is normalised in some national contexts.
— James Baker (@j_w_baker) June 25, 2018
.. international partners need to know how overheads work, differences between who can and cannot be hired.
T-AP is very resource intensive for the individual funding agencies.
Excited to be at "Where is the open in DH?" -- thanks, @gimenadelr @DanielPaulOD Robin Champieux for organizing! #dh2018
— Dr. Paige C. Morgan (@paigecmorgan) June 26, 2018
Robin Champieux Emphasis on open in many contexts (e.g. in North America) on increasing the impact of your work. This is problematic, not least because it is one-directional. But open scholarship more generally, includes not only open data / science / access / educational resources but equality/diversity/inclusion .. the wrong OA license can severely inhibit the reuse of data .. grassroots advocacy of the value of impact with respect to the community .. diversifying evaluation of published work is aligned with open principles.
Daniel O’Donnell First law of humanities computing:
The novel and non-trivial application of computation to humanities research problems inevitably requires an examination of first principles including the social, political, economic, and disciplinary rationale for the research itself.
Scientists know what they are doing: publishing papers in journals .. John Unsworth's "Scholarly Primitives" is so new compared to the scientific methods .. so if we don't share what our publication forums are, our data are, our methods, then being open in the humanities is hard .. don't be closed by accident.
Open in Latin America .. suffer from lack of resources and tech uptake .. OA needs to be global.
I tend to agree. At the risk of sounding overly blunt, I think it's a residual knee-jerk reaction against anything that sounds corporate/commercial.
— Dr. Paige C. Morgan (@paigecmorgan) June 26, 2018
I'm recalling how much I was cautioned early on not to talk about my DH project "b/c someone might steal it!" Terrible advice.
I think that NC is actually quite clearly defined (e.g. an OA article is certainly NC, an academic book distributed [=sold] through a publishing company is definitely not NC). I think that SA actually does much more harm in closing doors that authors meant to keep open.
— Walter Scholger (@wscholger) June 26, 2018
Brian Rosenblum Bottlenecks in the Open-Access System: Voices from Around the Globe. Access to reading is still an issue .. how do we decide where to invest our money/time in open access initiatives? Is APCs a model we should support as librarians.
...
Policy a key part of open access work: especially with government agencies around data. But you also need to put frameworks around this: ethical, legal.
#DH2018 Things I'm thinking in the 'Where is the Open in DH?' workshop: OA is just a start, OA always needs legal/ethical/policy frameworks, OA is advocacy, OA is both colonial and decolonial, OA in the humanities is always both an area of enquiry and an implementation.
— James Baker (@j_w_baker) June 26, 2018
OA in South America is bound up with problems of access to education, of lack of access to electricity.
The CC license does not override national copyright exceptions - if something can be reused through a legal exception - fair use or in continental European IPR legal"free license", e.g. for research and education - it cannot be restricted by a license. @cantavest #DH2018 https://t.co/Z0Zu70Jj12
— Walter Scholger (@wscholger) June 26, 2018
...
Paige Morgan Open source. There might be reasons to be concerned about MS buying GitHub, but MS owning your project isn't one: licences don't work that way. Lots of confusion out there: e.g. free software movement != open source software movement. Perceptions of open vary: development method, corporate ethos, political identity, economic choice. We have work to do to figure out how much we need to know and the locations in which we invest expertise. Peter Galison (1997) concept of "trading zones" important here. Strong desire to build everything ourselves so as to not be dependent on corporates or vendors, and yet rarely is the support for that development there. Privilege of the global north lowers the stakes of choices around openness: we have more resources if everything goes wrong.
Very relevant point to highlight when we talk about #openness in 'the North' and 'th South' #DH2018 👇🏽 pic.twitter.com/sHJn4rvlPO
— Ivonne Lujano (@ivonnelujano) June 26, 2018
David Wrisley Building Open Knowledge Foundation Definition of Open into work making available corpora of medieval French. Thinking through how to recognise participation/labour: various steps of the process captured in metadata as a form of open.
Ivonne Lujano Latin American community very convinced that research outputs should be publicly available, especially as subscription models are all but impossible to afford. System on incentives an important component of any open ecosystem: eg good things likeDORA, and bad things like Impact Factor.
+1, no, +10,, +100, to the DOAJ's work in helping to correct the often bad assumptions about OA journals propagated by B****'s list. #dh2018
— Dr. Paige C. Morgan (@paigecmorgan) June 26, 2018
Walter Scholger Geisteswissenschaftliches Asset Management System. Humanities data repository. Promotes open standards - eg TEI/XML - plus makes stylesheets developed available. Fear of others making a profit from your work creates desire for Creative Commons Non-Commercial. Multilingualism is hard.
Paula Ricaurte OpenLabs citizen innovation. Focus on open educational resources and open prototyping.
...
The keynote from @JanChvzSanti on digital language sustainability reminds me strongly of the work involved in encouraging the use of the #Welsh language in social networks and digital media at @SwanseaUni @Prif_Abertawe @AcademiHTeifi #digitalhumanities #DH2018 @GwennoFfrancon
— David M. Berry (@berrydm) June 26, 2018
¡ Vamos Mexico! ¡ Gracias Corea! 🇲🇽🇰🇷 #DH2018 pic.twitter.com/hZD8vkabsr
— Caitlin Christian-Lamb (@christianlamb) June 27, 2018
Micki Kaufmann, Data Viz and Interactive Storytelling for Diplomatic History ~18k letters/memos related to Henry Kissinger ..
man, @MickiKaufman is smart #dh2018
— Ethan Watrall (@captain_primate) June 27, 2018
.. using isometric viz shows how Kissinger shifted types of communication (x time, y topic, z volume of topics) .. difficult to unpick network graph 'hair balls': but you can unpack a hairball using 3d, rotating it, playing it back and rotating it.
@MickiKaufman showing really effective stacked bar graph visualisations from the 'Quantifying Kissinger' project: https://t.co/7CxFCU3j48 #DH2018
— Greta Franzini (@GretaFranzini) June 27, 2018
Love how 3D enables viewer to better see relationships w/o lines crossing. Plus also cheered with the charming metaphors of 'hairballs' and 'potatoes' - you can see all the bumps of the potato in 3D #DH2018 https://t.co/GOZA0Wzx3J
— Purdom Lindblad (@Purdom_L) June 27, 2018
I love so much about @MickiKaufman's work but am smitten right now with her articulation that you can resolve a "hairball" if you can move into additional dimensions. It's such an great way to describe why n-D work can be valuable #dh2018
— Jacqueline Wernimont (@profwernimont) June 27, 2018
Rachel Hendery & Kate Richards, Participatory creation of immersive First Peoples' spaces in Australia What are the challenges for people whose expectations of learning are very different from the dominant paradigm in Austrialian education?
Powerful acknowledging where research has occurred and presented - so often on Indigenous Lands. @RHendery sharing participatory immersive projects in collaborations with First Peoples of Australia #DH2018 #PS06
— Purdom Lindblad (@Purdom_L) June 27, 2018
Message stick with symbols familiar to First Peoples' students .. VR Fields of Dreams experience. Not able to fastforward videos, design choice that acknowledges respect for elders.
Powerful blend of physical objects and 3D immersive spaces, drawing on Indigenous Knowledge ways, mapping, and sharing information. @RHendery highlighting design decisions of Field of Dreams project. #DH2018 #PS06
— Purdom Lindblad (@Purdom_L) June 27, 2018
+ appreciate @RHendery real talk about funding, paying for labor, mixing/re-making for what you want given resources available. #DH2018 #PS06 https://t.co/eKkz1g71ml
— Purdom Lindblad (@Purdom_L) June 27, 2018
#dh2018 #PS06 @RHendery good on ethical challenges, competing aims re First Peoples of Australia. e.g. 1)Fields of Dreams: Message stick, VR of university experience. No-pause of videos of the Elders. 2) Generations of Knowledge video.3)Keeping Strong. No appropriation=no fund!
— Alison Booth (@alison_booth) June 27, 2018
Steven Jones, Reconstructing the First Humanities Computing Center Lost lab used by Roberto Rusa. Modelling absences from photography. Walter Benjamin 1928 "card index marks the conquest of three-dimensional writing". Using 3d modelling to see what we can't see in written accounts. Lack of humans from the 3d model of a research workspace reminds us that humans are the blackbox.
#dh2018 #ps06 @s3jones reconstructing Busa, speculative media archeology, N-dimension modeling. Benjamin on the card catalogue leads to VR of the first DH lab, its machines.
— Alison Booth (@alison_booth) June 27, 2018
differences between archival photos (single operators) vs personal photos from operators (pair programming) show two different modes of working, pairs hidden in archival record of the First Humanities Computing Center @s3jones #DH2018
— Purdom Lindblad (@Purdom_L) June 27, 2018
Geoffrey Rockwell, Experiments with alternative and virtual reality games Serious games to bridge between parts of the city: Zombie game to help people orientate the city .. when working with 'real people' they have relatively little patience when the technology of finicky ..
after presenting compelling game work, @GeoffRockwell closes with sharing failures (which is so helpful) - difficulties communicating point of game, finicky but improving tech, small interface (usability ?s), + ppl dont want to install stuff. #PS06 #DH2018
— Purdom Lindblad (@Purdom_L) June 27, 2018
What a great array of projects using 3D technologies. Still seems web GL is most useful for providing public access. And fAR-play for augmented reality looks neat for phone-based applications. Also appreciate the acknowledgment of the many challenges involved #DH2018 #PS06 pic.twitter.com/YVAR8jfGTv
— Jasmine Mulliken (@jasminemulliken) June 27, 2018
James William Baker, Ben Jackson, David Berry, Sharon Webb, Rebecca Wright {our session so no tweeting!}
Challenges in Enabling Mixed Media Scholarly Research with Multi-media Data in a Sustainable Infrastructure - Roeland Ordelman, Carlos Martínez Ortíz, Liliana Melgar Estrada, Marijn Koolen, Jaap Blom, Willem Melder, Jasmijn Van Gorp, Victor De Boer, Themistoklis Karavellas4, Lora Aroyo, Thomas Poell, Norah Karrouche, Eva Baaren, Johannes Wassenaar, Julia Noordegraaf, Oana Inel CLARIAH (Common Lab Research Infrastructure for the Humanities) project. Supporting Dutch research. WP5 looking at media studies: newspaper, film, audio. Research pilots to test infrastructure: e.g., cross media research on drugs .. 'metadata archeology' required to understand what metadata fields mean and document them .. fecerated login to use restricted (e.g. copyright) materials .. time-based media, auto metadata extraction using speech extraction .. CLARIAH Media Suite .. using Scholarly Primitives as main guidance for development
#DH2018 #LP07 Unsurprisingly given the good people involved, @CLARIAH_NL Media Suite looks robust and well thought through https://t.co/32Wn51LqTU Interesting that it includes lots of open data, but that tools to work with that data require login/membership.
— James Baker (@j_w_baker) June 28, 2018
Jupyter notebook integration: generic viz as a way into the data, Jupyter as the next step.
#DH2018 #LP07 @RoelandOrdelman says the @ProjectJupyter integration with @CLARIAH_NL is *very* new. I love Jupyter, especially as a powerful pedagogical tool. I'd be curious to see if the integration here is used to teach data intensive humanities research in NL.
— James Baker (@j_w_baker) June 28, 2018
Increible presentación de la Media Suite de CLARIAH. Estoy muy emocionado poder explorar el potencial para el analisis de la producción multimodal en las plataformas sociales! #dh2018 #LP07
— Hans Bouchard (@HansRUB) June 28, 2018
Data Scopes: towards Transparent Data Research in Digital Humanities - Rik Hoekstra, Marijn Koolen, Marijke van Faassen DECK. Stepped away from the macroscope project: see it as making the data something to look at, rather than integrated into the research process .. Example here discourse of migration 1950s-1970s: shift to neoliberal dominance within discussions of migration .. research process (van Faassen & Hoekstra 2017) .. computational approach requires modelling data (McCarty 2004) .. limited number of separate activities working that are part of this research process that produces the data scope: Modelling represents the data in such a way that it will fit the research scope; Selection chooses datasets and parts of datasets that are relevant for the research questions; Normalizing structures data and reduces data variation so that they may be queried more easily and they can be used for comparisons, classifications or calculations; Linking data connects previously unconnected data, providing them with context from other data sets. Researchers should be aware that the validity or relevance of links can be context-dependent (person X and Y are linked for a specific question because they played the same role, but the link may be invalid for other questions, see Brenninkmeijer et al 2012); Classifying data groups them in in order to reduce complexity. This adds a level of abstraction to the data .. capturing what you did and what you didn't know, helps convey your research: not just results, but also publishing process as part of the real research that goes into results. You always make choices in research!
Really digging 'Data scopes: Towards transparent data research in digital humanities.' Scoping your data is not 'mere preparation'—it's a set of core scholarly design choices of any research project using data. #DH2018
— Chelcie Juliet Rowell (@ararebit) June 28, 2018
https://t.co/XaSexhsGxu #DH2018
A Deep Gazetteer of Time Periods - Ryan Shaw, Adam Rabinowitz, Patrick Golden PeriodO. A gazateer of time periods. Capture history and development of periodisation in different fields. Many terms - e.g. "Archaic" period - ambiguously used: 35 different "Archaic" periods in the gazeteer. Entries must have names, temporal bounds, citable source.
https://t.co/pPgpkj3jvw requires name, temporal bounds, geographic region, citable source. Space & time inextricably linked. #SpatialHumanities #HGIS #DH2018 #LP07 https://t.co/ohPdq0VPDH
— Hannah L Jacobs (@hannahljnc) June 28, 2018
#DH2018 #LP07 @perio_do: a gazetteer for time periods https://t.co/eG88yQdgc0 I really like the idea. But it does seem at present heavily skewed to archaeological definitions: e.g., see 18th century for which there is no 'Long Eighteenth Century'. Hope that changes.
— James Baker (@j_w_baker) June 28, 2018
Citable sources are looking into WorldCat and VIAF. Uses ARK identifiers. Changes and additions to the database can be proposed.
Projects like @perio_do make me want to run away to join a PhD program in information science. This is the disciplinary expertise of information organization, and it's my jam. #DH2018 #LP07
— Chelcie Juliet Rowell (@ararebit) June 28, 2018
Q&A: what to do with politically contested time periods.
#DH2018 #LP07 Great question/comment from @ecbmurphy ~ in literary studies, as time periods get more modern the geographical region of the time period is much less relevant.
— James Baker (@j_w_baker) June 28, 2018
#DH2018 #LP07 Another great question (don't know who asked it) about how to model politically contested time periods and to ensure the data is decolonial and doesn't enforce privileged positions.
— James Baker (@j_w_baker) June 28, 2018
Creating and Implementing an Ontology of Documents and Texts - Peter Robinson A text is an instance of an act of communication inscribed in a document.
Perhaps like this? #dh2018 pic.twitter.com/Qg774Zj5IA
— M. H. Beⓐls (@mhbeals) June 28, 2018
A machine learning methodology to analyze 3D digital models of cultural heritage objects - Diego Jimenez-Badillo, Salvador Ruiz-Correa, Mario Canul-Ku, Rogelio Hasimoto Using geometric form as a search 'term'. 10-15k masks from Mexico. Question: how did one style of mask develop into the next? Answer: produce artificial masks to bridge between styles.
#DH2018 #LP10 Diego Jimenez-Badillo cracking out the counterfactuals/imaginaries to understand the evolution of 3d heritage objects ~ Question: how do we understand how one style of mask developed into the next? Answer: produce artificial masks to bridge between styles.
— James Baker (@j_w_baker) June 28, 2018
#DH2018 #LP10 "A Machine Learning Methodology To Analyze 3D Digital Models Of Cultural Heritage Objects". Must admit, I'm smitten by this paper. https://t.co/0ngEpkHipF pic.twitter.com/jo4hPAzJY4
— James Baker (@j_w_baker) June 28, 2018
Deformation model. But deep learning. Computationally intense.
"If we had a better machine we could run it faster" Get this man a HPC! These masks are amazing #dh2018
— M. H. Beⓐls (@mhbeals) June 28, 2018
Data outputs that record/represent intensity of change.
#DH2018 #LP10 Figuring out how one mask style got to another using deformation modelling. Super cool! https://t.co/0ngEpkHipF pic.twitter.com/aPouJpe50d
— James Baker (@j_w_baker) June 28, 2018
Interpreting Difference among Transcripts - Michael Sperberg-McQueen, Claus Huitfeldt Deck My comparing different transcripts of the same exemplar, we find out things about transcription more broadly. Transcript: all the tokens in the transcript and the exemplar are 1 to 1, the first instantiates the latter. Every transcript represents a reading of an exemplar. Purpose of a transcript determines what tokens are transcribed. Transcribes often replace one type with another is a formalised way (e.g. to represent italics). From any transcription we can reconstruct the reading of the exemplar (as well as the text).
...
<script async src="https://platform.twitter.com/widgets.js" charset="utf-8"></script>So, why an American University have bought a Colombian archive? Because Colombian government neglected archives preservation. This is the digital repository in progress doing by Vanderbilt University and share to us by Eduard Arriaga #DH2018 https://t.co/PCwu7MGb1g
— Jairo Melo (@jairoamelo) June 28, 2018
...
The (Digital) Space Between: Notes on Art History and Machine Vision Learning - Benjamin Zweig
Rogues Gallery project. Understand the difference between curatorial categorisation/description of artworks and machine (Microsoft Cognitive Services) tagging. Odd tags. What assumptions being made around machine vision learning? And how do these differ to what art historians do. Machine vision learning looking for difference in subject matter, colour, line, et cetera: this is formalism in art history - Wolfflin, Panosky et al. But for art historians this is a minefield: machine vision learning presents a narrow and outmoded conception of art history, Western assumptions, misrepresentation of formalism as the only kind of art history .. perhaps - however - we need to change our expectations, ask what the machine can do rather than worry about what it can't .. Alois Riegl as an alternative? Stifragen, on discernable grammar of ornaments. Here we can use machine vision to solve real problems (lots of stuff) in a global way. Finding ornamental patterns in unexpected places.
[I love that @medievalben is offering alternate intellectual genealogies for digital art history. Ok, art historians know Riegl, but he's not a reference point for outsiders] #DH2018
— Matthew Lincoln (@matthewdlincoln) June 29, 2018
Modeling the Fragmented Archive: A Missing Data Case Study from Provenance Research - Matthew Lincoln, Sandra van Ginhoven
Database approaches to the history of the art market. Data for the dealer Knoedler. Problem of incomplete data: e.g. incomplete date info (year/month but no day, month/day but no year). Communicate techniques for estimating datapoints to our audience in ways they understand: e.g. art historians don't understand what 'confident intervals' mean. Indeed, art historians infer and make educated guesses for missing data all the time.
@matthewdlincoln: how can we use conventional methods for missing data in our history of art markets & communicate them to our audience? #dh2018
— Emily Esten (@sheishistoric) June 29, 2018
For example, use stock book numbers to guess at precise date when data is missing. 20-30k artworks. Every time we make a new guess, rerun the whole analytical flow. Real research question: is History painting a risky genre to invest in? With only correct data, yes. With estimates in, no. But presenting this as a box plot doesn't make sense to a historian, causes suspicion.
#DH2018 #SP15 @matthewdlincoln and his colleagues at the Getty Research Institute are such thoughtful digital art historians: rather that demanding art historians learn box plots, they are developing alternative data viz that communicates more effectively to the community. https://t.co/6EyGag8Eis
— James Baker (@j_w_baker) June 29, 2018
Take homes: make your model intelligible, make multiple viz, reduce suspicions.
Another elegant presentation by @matthewdlincoln : how to communicate stats-driven data to art historians who don’t know or trust “multiple imputation” or “confidence intervals” through animated visualizations #dh2018 https://t.co/xfQr85Fmdy
— Michelle Dalmau (@mdalmau) June 29, 2018
@DukeU grad Sandra van Ginhoven & @matthewdlincoln using simulations, literal graphic fuzziness to show possible results. @GettyHub #DH2018 #SP15
— Hannah L Jacobs (@hannahljnc) June 29, 2018
It made feel slightly (only slightly) less like a loser :-) when I saw that @matthewdlincoln and colleagues worked with something like 6 IT specialists for implementing and running the data models and visualizations. It takes a village (when you have one). #dh2018
— Michelle Dalmau (@mdalmau) June 29, 2018
Urban Art in a Digital Context: A Computer-Based Evaluation of Street Art and Graffiti Writing - Sabine Lang, Björn Ommer
How does urban art benefit from digital possibilities? Urban art is context dependent. Illegality means that it is often removed, destroyed. So digitisation is important to survival. But also analysis, given the scattered nature of the genre. Visual search to track the reception and spread of a motif, as well as its relation to urban context. Process using machine vision followed by human feedback into the model.
Extracting and Aligning Artist Names in Digitized Art Historical Archives - Benoit Seguin, Lia Costiner, Isabella di Lenardo, Frédéric Kaplan
Pipeline from archive to database. Recto-verso digitisation machine built for the project. 1,500 images per day digitised.
.@Seguin_Be discussing his dissertation work on end-to-end automated metadata generation pipeline for photo archives #DH2018
— Matthew Lincoln (@matthewdlincoln) June 29, 2018
Limited data per image: how to connect this to existing data. Challenges: name variations, substitutions, implicit knowledge, notation schemes for copies.
Like a good engineer, @Seguin_Be was thoughtful about generalizing the pipeline so that it could work for different kinds of historical documents, not just the ones he was working on. #DH2018
— Matthew Lincoln (@matthewdlincoln) June 29, 2018
Use computer vision to find conflicting attributions in the database.
dhSegment tool for processing documents.
@Seguin_Be mentions dhsegment as software applied across their collection to segment part of images - see https://t.co/NEjkEtnBl1 and code at https://t.co/wAa5objSuk - We should try it out #DH2018 #SP15
— Arianna Ciula (@ariciula) June 29, 2018
Métodos digitales para el estudio de la fotografía compartida. Una aproximación distante a tres ciudades iberoamericanas en Instagram. - Gabriela Elisa Sued
Digital methods for the study of shared photography. A distant approach to three Ibero-American cities on Instagram.. DECK. Shared urban photography. 5k pictures published in October 2015 with city related hashtags. Research questions:
- What kind of aesthetic and thematic representation patterns are observed in each of the labels?
- Is photography sharing on Instagram a community practice or can it only be viewed as a scattered collection of photographs?
- What social uses are given to urban labels, beyond being mere organizers of photographs on cities?
- What kind of citizen experiences or ways of living in the city are identified in the photographs studied?
- Can stable elements be identified in the thematic, aesthetic and community aspects of the practice that allows thinking about the formation of a discursive genre of its own?
Using a combination of textual, visual, and network analysis.
@gabysued studies shared photography practices on @instagram using impressive range of analytical techniques: content, textual, visual, network, statistical. Recognizing the dialogs that can form across techniques. #DH2018 #SP15
— Hannah L Jacobs (@hannahljnc) June 29, 2018
High level of thematic recurrences within the photographs in the dataset. Visual clusters on images that get little attention: sunrises, sunsets.
Lorena Gauthereau, Hannah Alpert-Abrams, Alex Galarza, Mario H. Ramirez, Crystal Andrea Felima
Now up - Revisiting the Archivo Mesoamericano: Digitization and the Revolutionary Histories of Central America and Mexico / Repasando el Archivo Mesoamericano: digitalización y las historias revolucionarias de Centroamérica y Méxcio @mario_h_ramirez #ps25 #dh2018
— bibliotekah (@tttkay) June 29, 2018
Like many post-custodial projects, one of the goals of @mario_h_ramirez work is to build capacity across institutions, and especially with partner institutions in central america...
— Hannah Alpert-Abrams (@hralperta) June 29, 2018
...but this costs $$$#dh2018 #ps25 https://t.co/4Q4cb7p9UC
Eternal Sunrise: digital life-cycles and long-term preservation for social justice, Hannah Alpert-Adams [DECK}(https://drive.google.com/file/d/1aMdxHtg-6Rmd-3L8Cu7UFtTQ5mqj2fWa/view) Police records 1882-1996, 80 million pages, 11 million digitsed
I'm talking about the history and future of the @llilasbenson digital portal to the Guatemala Police Archive: https://t.co/IONdBzkoOB
— Hannah Alpert-Abrams (@hralperta) June 29, 2018
Slides in english and spanish are available here; open the google doc for some useful links:https://t.co/NM6bYkmEPh#dh2018 #PS25
11 million documents put online in under a year: Archivo Histórico de la Policía Nacional de Guatemala. Argument: The conditions of social justice & human rights projects make them uniquely vulnerable to long-term planning challenges. Why? Ethical complexity; urgency to get them up can make you forget long term planning; system poor at handling labour, inequity, trauma .. Working under urgency requires documentation and not just taking people with the best skills: also need people with cultural sensitivity/knowledge .. Minimal Computing as a design choice to enable access (especially as archive used in lega processes)
Ethical challenges of long term preservation of human rights collections in academic institutions using Guatemala Historical National Police Archive (AHPN) @hralperta #ps25 #dh2018
— bibliotekah (@tttkay) June 29, 2018
Hannah Alpert-Abrams on the importance of minimal computing to ensure accessibility beyond academic publics and global north #DH2018 #PS25 https://t.co/b5bWW3Xu8t
— Felicity Tayler (@ftayler) June 29, 2018
The GAM Digital Archive Project - Alex Galarza Site .. ensuring chain of custody vital when archive used in human rights trails ..
Haverford brought scanners to Grupo Apoyo Mutuo (GAM) to digitize their documentation on disappearances; GAM hopes the documentation can be used for legal justice @galarzaalex #ps25 #dh2018
— bibliotekah (@tttkay) June 29, 2018
Metadata standards used balanced standards with practicalities that emerged from the collections and social context: e.g. aimed to take into account younger Guatemalans for whom the lived experience of the history behind the documents was absent.
#DH2018 #PS25 papers beautifully model how we should be working with communities and centering how communities will use digital archives. Metadata, organizational schemes come from these communities.
— Dr. Anelise H. Shrout (@AneliseHShrout) June 29, 2018
Methodology of the Oppressed: Approaching US Latina/o DH through Decolonism and Affect - Lorena Gauthereau 'Recovery': collecting, preserving, and disseminating latina/o literature to correct their under-representation in archival holdings. A cognisant de-colonial archival methodology views items in collections as belonging to the community.
"Migrant archives reside in obscurity and are always at the edge of annihilation." Love @LGauth19 use of Rodrigo Lazo's framing of migrant archives in #DH2018 #ps25
— Hannah Alpert-Abrams (@hralperta) June 29, 2018
Post-colonial critiques look at existing power. De-colonial work functions at a micro level (e.g. Franz Fanon, The Wretched of the Earth) - producing something new.
Perez, Emma. “Queering the Borderlands: The Challenges of Excavating the Invisible and Unheard.” Frontiers: A Journal of Women Studies 24, no. 2 (2003): 122–31. https://doi.org/10.1353/fro.2004.0021.
Teaching beyond the Haitian Exceptionalism through Digital Decolonization and Social Justice Pedagogy - Crystal Andrea Felima
Works as liaison to @dLoCaribbean and teaches on Haitian Studies at @UF informed by critical pedagogy, black feminist thought and social justice @phelima #ps25 #dh2018
— bibliotekah (@tttkay) June 29, 2018
Haitian's subject to narratives of being superhuman (they survive disasters) and subhuman (they are corrupt). Showing alternative images of Haitians as a form of decolonial social justice pedagogy. Positionality of Haiti within DH that it is often not Haitians working on those projects. Outputs as student project: e.g. Gender in Haiti
Love how @phelima encouraged her students to use @uflib resources in their research. #postdocsinthelibrary #PS25 #DH2018 https://t.co/bhfWAIIt1I
— Patricia Hswe (@pmhswe) June 29, 2018
#ps25 @phelima's teaching philosophy is v inspiring. all the questions she raises around labor, language, agency are super relevant to archives/archivists and how we present the collections we steward esp related to marginalized communities #dh2018
— bibliotekah (@tttkay) June 29, 2018
Q&A
Libs ethos in the States is not to centralize content. We share all the time to @dpla & other aggregators. This gets to a point @hralperta was making about combating the uni-directional post-custodial model. We need to share back to repos and portals based in LatAm&Carib #dh2018 https://t.co/ZD2dcgw35P
— Michelle Dalmau (@mdalmau) June 29, 2018
Precarity of this work compared to funding available for things like Shakespeare's first folio .. digitisation creates a second archive, another thing to look after .. making it ourselves whether or not we are ready is resistance .. who is licensed to have an opinion about how something is described?
Audience question: do we have the capacity to build postcustodial partnerships? Are we ready?@LGauth19: If we ask that, we might never get the work done. The world was not made for marginalized groups. We need to do the work. #ps25 #dh2018
— bibliotekah (@tttkay) June 29, 2018
Things I've taken from #DH2018 #PS25 "Social Justice, Data Curation, and Latin American & Caribbean Studies": things to read (Perez, Queering the Borderlands), urgency demands extra documentation, collections must belong to communities, metadata standards need to be poco hacked.
— James Baker (@j_w_baker) June 29, 2018
@LGauth19 speaks about labor of @APPRecovery as postcustodial model. Minorities don't have the luxury of waiting until institutions come up with best practices for this, we must use the "master's tools to dismantle the house" #audrelorde #usldh #chingona #ps25 #dh2018
— Gabi Baeza Ventura (@gbaezaventura) June 29, 2018
Anelise Hanson Shrout, Jamila Moore-Pewu, Gimena del Rio Riande, Susanna Allés, Kajsa Hallberg Adu
First-Gen DH in the United States - Anelise Hanson Shrout Deck. Most DH pedagogy discussion Anglophone. DH can undertake the digital divide and combat structured inequality. In the US: distribution of DH is around elite institutions, first generation students are heavily represented outside elite institutions. Work here to reshape pedagogy around first-gen students. These students are: better at collaboration, experts in mobile tech, already using digital tools creatively.
@AneliseHShrout:
— Emily Esten (@sheishistoric) June 29, 2018
First-gen students are
- better at collaboration
- experts with mobile technology
- already use digital tools for humanistic questions#dh2018
So, take really seriously the DHish skills that they already bring: e.g. two minute script for a podcast/broadcast using a project from their lives.
@AneliseHShrout speaking on decolonising digital pedagogy: we design assignments that introduce students to DH and introduce them to new ways of looking at themselves and their communities. #ps28 #dh2018
— Alex Turton (@AlexanderTurton) June 29, 2018
Teaching DH: The Other Linguistic Divide - Susanna Allés Modern language departments haven't integrated DH in the same ways that English departments have. One challenge is the lack of fluency of students with the language they are learning.
Susanna Allés-Torrent highlights Anglophone #DH transculturation as challenge to #DHPedagogy in modern language courses #DH2018 #PS28 #PS30
— Hannah L Jacobs (@hannahljnc) June 29, 2018
In the context of modern languages, goal of using digital tools to analyse literary texts is to give the students more confidence to write about authors writing in the language they are learning.
Allés-Torrent used #textanalysis, @VoyantTools, Antconc, @rstudio for distant reading of Spanish texts that also contributed to vocab & comprehension development. #DH2018 #PS28 #PS30 https://t.co/0q244qmVBB
— Hannah L Jacobs (@hannahljnc) June 29, 2018
Teaching DH "From Scratch" - Gimena del Rio Riande Never truly from scratch, someone has already done something previously. In Argentina circa 2011, one challenge was building materials and documentation in Spanish.
@gimenadelr discusses long history of #HumanidadesDigitales in Argentina through Incipit Journal https://t.co/Eh7v8sVIBu #DH2018 #PS28
— Hannah L Jacobs (@hannahljnc) June 29, 2018
Lots of interest in open science, questions of access, and building knowledge. Repurposing software for pedagogical purposes. We have communities of practice and no clear audience.
@gimenadelr notes that orgs like @aahdArg @Red_HD have led #DHPedagogy in their communities through workshops. Greatly benefit from audiences being not only students but also faculty, staff, librarians. #DH2018 #PS28 #PS30
— Hannah L Jacobs (@hannahljnc) June 29, 2018
DH Pedagogy from the Global South & Ghana - Kajsa Hallberg Adu Pan-African student body .. decolonial pedagogy takes account of different knowledges (indigenous thought) and different styles of learning (the work of groups)
.@kajasaha: references hooks(2009) and Freire (2000) #DH2018 #PS28
— Lorena Gauthereau (@LGauth19) June 29, 2018
Freire opposes the idea that students are no knowledgable .. use of WhatsApp (a known space) as a lecture forum .. Wikimedia work, supported by the Foundation ..
.@kajsaha discusses decolonial thought and engaged pedagogy and how this applies to developing DH at @Ashesi (and more widely, the global south) #DH2018 #PS28 pic.twitter.com/bxKRu5d94o
— Caitlin Christian-Lamb (@christianlamb) June 29, 2018
.. important to challenge traditional texts produced in the Global North as the only sources of knowledge and thereby make it clear to students that they can create knowledge
Megan Finn Senseney, Paige Morgan, Miriam Posner, Andrea Thomer, Helene Williams Panel Abstract
Megan Finn Senseney We cannot teach DH without humanities data .. so the panel is about the various ways we get hold of the data .. we need to articulate the features of a good pedagogical dataset .. choosing a good dataset for teaching: does it fit my uses? is it the right size for my course? is there enough context and provenance? is it documented? is it messy enough but not too much? .. if we are too anxious to share untidy data we won't have any good pedagogical data .. minimal viable curation: plenty of provenance, but too much cleaning
Learning from our mistakes: using legacy projects to create better librarian/faculty collaborations - Paige Morgan Found all these legacy projects supported by the library: was allowed to fix them and wasn't allowed to take them down ..
@paigecmorgan notes that the best way to get started in digital scholarship is to see the slimy [infrastructural] under belly #DH2018 #DH_PS32 #PS32
— Hannah L Jacobs (@hannahljnc) June 29, 2018
.. legacy websites are more complex challenges that 'is it broken or not'?: sometimes things are dated in their view of the world (e.g. colonial) .. Dead site examples: bibliography, migrated it to GitHub, credited original work .. this is static preservation, but what might we aim for beyond that?
Can we Please Have Your Old Data? - Miriam Posner Deck. Ideal size of records: ~2k records - so much data that it is inconvenient to change everything but hand, but small enough to not break excel so you can glance at data; good to have some change over time .. data needs to be the kind of data you can ask humanities questions of .. students have to do a data critique: what is and isn't mentioned.
DH in MLIS - Helene Williams Librarians need to be explain to a humanities collaborator why a dataset might not be what they need.
Librarians need to know how to find, evaluate, and manipulate data across disciplines; and they need to be able to frame/contextualize that dataset for researchers, says @bibliorogue #DH2018
— Dr. Paige C. Morgan (@paigecmorgan) June 29, 2018
"@bibliorogue's final project for students is to look for a real DH site that has problems and take on the role of librarian to improve it #DH2018" #DH_PS32 https://t.co/oBzZRMeIeg
— Lawrence Evalyn (@LawrenceEvalyn) June 29, 2018
Great project idea from @bibliorogue for a DH for Librarians class: Find an aging DH project and come up with a plan for adapting or preserving it. #dh2018 #PS32 pic.twitter.com/5Su5a5yAzt
— Miriam Posner (@miriamkp) June 29, 2018
Start small!
Megan Finn Senseney Data salting: strategic insertion of errors in data.
.@modernmuchness: how are you supposed to teach data curation when all of the available data is already curated? #DH2018 #DH_PS32 🤔 pic.twitter.com/KuSzGdiQ7L
— Miriam Posner (@miriamkp) June 29, 2018
Data with good provenance and contextual information materials are useful for catalysing discussions of hidden/disguised labour in curatorial work.
#DH_PS32 Q: What readings to assign in DH classes? -- Against Cleaning https://t.co/qt7f6aO631 , You Say Data I Say System https://t.co/W9GXlxKPaM#DH2018
— Lawrence Evalyn (@LawrenceEvalyn) June 29, 2018
Closing Keynote: Digital Experimentation, Courageous Citizenship and Caribbean Futurism - Schuyler K Esprit
Founder of Create Caribbean: first DH Lab in the Caribbean (Dominica). Destroyed by Hurricane Maria.
.@elotroalex: @CreateCaribbean proves that a DH lab can be a conduit for very young individuals to build their past and their future. #dh2018
— Élika Ortega (@elikaortega) June 29, 2018
Maria was a climate chance disaster.
.@schuyleresprit: "How do we understand community in the era of climate disaster?" #dh2018
— Christina B (@clboyles) June 29, 2018
The extent of destruction in Mahut by Hurricane Maria brought up questions of home and homelessness. These anxieties shaped the Carisealand project. @schuyleresprit #DH2018
— Rebecca Munson (@Shxperienced) June 29, 2018
How can the performing arts contribute to climate policy change? [Casiseland[(http://carisealand.org/) project: looks at links like Foreign Aid and "Disaster Porn" complexes, environment law and cultural heritage.
A good set of topics for thinking through climate change intervention effectively. #DH2018 pic.twitter.com/zUr4XjFARb
— Dr. Paige C. Morgan (@paigecmorgan) June 29, 2018
Imagining alternative futures not just restoration in the face of climate violence.
Inspiring call to center climate change at center of questions of the humanities - @schuylerespirt #DH2018 https://t.co/DcKDKdg6Fu
— Purdom Lindblad (@Purdom_L) June 29, 2018
With mobile penetration much higher than broadband in many parts of the world, digital work that seeks to build communities and drive rebuilding must be mobile friendly in user interface and hardware resources@schuyleresprit#DH2018
— Lisa Tagliaferri (@lisaironcutter) June 29, 2018
#DH2018 @schuyleresprit mentions that mobile penetration is much higher than broadband in Dominica - especially in the wake of climate catastrophe - meaning that stuff needs to work on mobile. So, for the second time today, Minimal Computing is important! https://t.co/HhL4MTijDH
— James Baker (@j_w_baker) June 29, 2018
Principles of rebuilding have to fit with how people live, love, and work.
.@schuyleresprit Principles of rebuilding include: Love (family histories, faith-based orgs, reduction of social problem); Live (land use, home ownership, food/water security, waste mgmt); Work (sustainable work, employment skills, retention) #DH2018
— Ashley SandersGarcia (@throughthe_veil) June 29, 2018
Making a contribution to ensuring that people know what they lose if they sacrifice to the needs of rebuilding to protect against the next storm.
So fascinating the way @schuyleresprit and @CreateCaribbean are thinking about preservation of both physical structures and the stories they hold, in the Caribbean. #DH2018 pic.twitter.com/vqj8UECIsU
— ravon (@afroxmericana) June 29, 2018
Urge to not separate work of the Lab from the pressing concerns of the community.
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
Exceptions: embeds to and from external sources, and direct quotations from speakers