Page 215

What seems a clear line between publications and data in the sciences and social sciences is a decidedly fuzzy one in the humanities. Publications and other documents are central sources of data to humanists. … Data sources for the humanities are innumerable. Almost any document, physical artifact, or record of human activity can be used to study culture. Humanities scholars value new approaches, and recognizing something as a source of data (e.g., high school yearbooks, cookbooks, or wear patterns in the floor of public places) can be an act of scholarship. Discovering heretofore unknown treasures buried in the world's archives is particularly newsworthy. … It is impossible to inventory, much less digitize, all the data that might be useful scholarship communities. Also distinctive about humanities data is their dispersion and separation from context. Cultural artifacts are bought and sold, looted in wars, and relocated to museums and private collections. International agreements on the repatriation of cultural objects now prevent many items from being exported, but items that were exported decades or centuries ago are unlikely to return to their original site. … Digitizing cultural records and artifacts make them more malleable and mutable, which creates interesting possibilities for analyzing, contextualizing, and recombining objects. Yet digitizing objects separates them from the origins, exacerbating humanists’ problems in maintaining the context. Removing text from its physical embodiment in a fixed object may delete features that are important to researchers, such as line and page breaks, fonts, illustrations, choices of paper, bindings, and marginalia. Scholars frequently would like to compare such features in multiple additions or copies.

Page 214

Borgman on information artifacts and communities:

Artifacts in the humanities differ from those of the sciences and social sciences in several respects. Humanist use the largest array of information sources, and as a consequence, the station between documents and data is the least clear. They also have a greater number of audiences for the data and the products of the research. Whereas scientific findings usually must be translated for a general audience, humanities findings often are directly accessible and of immediate interest to the general public.

Page 204

Borgman on the different types of data in the social sciences:

Data in the social sciences fall into two general categories. The first is data collected by researchers through experiments, interviews, surveys, observations, or similar names, analogous to scientific methods. … the second category is data collected by other people or institutions, usually for purposes other than research.

Page 202

Borgman on information artifacts in the social sciences

like the sciences, the social sciences create and use minimal information. Yet they differ in the sources of the data. While almost all scientific data are created by for scientific purposes, a significant portion of social scientific data consists of records credit for other purposes, by other parties.

Borgman, Christine L. 2007. Scholarship in the Digital Age: Information, Infrastructure, and the Internet. Cambridge, Mass: MIT Press.

My notes

Page 147

Borgman on the challenges facing the humanities in the age of Big Data:

Text and data mining offer similar Grand challenges in the humanities and social sciences. Gregory crane provide some answers to the question what do you do with a million books? Two obvious answers include the extraction of information about people, places, and events, and machine translation between languages. As digital libraries of books grow through scanning avert such as Google print, the open content Alliance, million books project, and comparable projects in Europe and China, and as more books are published in digital form technical advances in data description, and now it says, and verification are essential. These large collections differ from earlier, smaller after it's on several Dimensions. They are much larger in scale, the content is more heterogenous in topic and language, the granularity creases when individual words can be tagged and they were noisy then there well curated predecessors, and their audiences more diverse, reaching the general public in addition to the scholarly community. Computer scientists are working jointly with humanist, language, and other demands specialist to pars tax, extract named entities in places, I meant optical character recognition techniques counter and Advance the state of art of information retrieval.

Page 137

Borgman discusses hear the case of NASA which lost the original video recording of the first moon landing in 1969. Backups exist, apparently, but they are lower quality than the originals.

Page 122

Here Borgman suggest that there is some confusion or lack of overlap between the words that humanist and social scientists use in distinguishing types of information from the language used to describe data.

Humanist and social scientists frequently distinguish between primary and secondary information based on the degree of analysis. Yet this ordering sometimes conflates data sources, and resorces, as exemplified by a report that distinguishes quote primary resources, ed books quote from quote secondary resources, Ed catalogs quote. Resorts is also categorized as primary wear sensor data AMA numerical data and filled notebooks, all of which would be considered data in The Sciences. But rarely would book cover conference proceedings, and he sees that the report categorizes as primary resources be considered data, except when used for text or data mining purposes. Catalogs, subject indices, citation index is, search engines, and web portals were classified as secondary resources.

Pages 119 and 120

Here Borgman discusses the various definitions of data showing them working across the fields

the following definition of data is widely accepted in this context: AT&T portable representation of information in a formalized manner suitable for communication, interpretation, or processing. Examples of data include a sequence of bits, a table of numbers, the characters on a page, recording of sounds made by a person speaking Ori moon rocks specimen. Definitions of data often arise from Individual disciplines, but can apply to data used in science, technology, the social sciences, and the humanities: data are facts, numbers, letters, and symbols that describe an object, idea, condition, situation, or other factors.... Terms data and facts are treated interchangeably, as is the case in legal context. Sources of data includes observations, complications, experiment, and record-keeping. Observational data include weather measurements... And attitude surveys... Or involve multiple places and times. Computational data result from executing a computer model or simulation.... experimental data include results from laboratory studies such as measurements of chemical reactions or from field experiments such as controlled Behavioral Studies.... records of government, business, and public and private life also yield useful data for scientific, social scientific, and humanistic research.

Pages 117 to 1:19

Here Borgman discusses the ability to go back and forth between data and reports on data she cites Phil born 2005 on this for a while medicine. She also discusses how in the pre-digital error data was understood as a support mechanism for final publication and as a result was allowed to deteriorate or be destroyed after the Publications upon which they were based appeared.

Page 115

Borgman makes the point here that while there is a Commons in the infrastructure of scholarly publishing there is less of a Commons in the infrastructure 4 data across disciplines.

The infrastructure of scholarly publishing Bridges disciplines: every field produces Journal articles, conference papers, and books albeit in differing ratios. Libraries select, collect organize and make accessible publications of all types, from all fields. No comparable infrastructure exists for data. A few Fields have major mechanisms for publishing data in repositories. Some fields are in the stage of developing standards and practices to activate their data resorces and Nathan were widely accessible. In most Fields, especially Outside The Sciences, data practices remain local idiosyncratic, and oriented to current usage rather than preservation operation, and access. Most data collections Dash where they exist Dash are managed by individual agencies within disciplines, rather than by libraries are archives. Data managers usually are trained within the disciplines they serve. Only a few degree programs and information studies include courses on data management. The lack of infrastructure for data amplifies the discontinuities in scholarly publishing despite common concerns, independent debates continue about access to Publications and data.

Page 41

discussions of digital scholarship tend to distinguish implicitly or explicitly between data and documents. Some of you data and documents as a Continuum rather than a dichotomy in this sense data such as numbers images and observations are the initial products of research, and Publications are the final products that set research findings in context.

A great paragraph here on the value of interconnection

scholarly data and documents are of most value when they are interconnected rather than independent. The outcomes of a research project could be understood most fully if it were possible to trace an important finding from a grant proposal, to data collection, to a data set, to its publication, to its subsequent review and comment period journal articles are more valuable if one can jump directly from the article to those insights into later articles that cite the source article. Articles are even more valuable if they provide links to data on which they are based. Some of these capabilities already are available, but their expansion depends more on the consistency of the data description, access arrangements, and intellectual property agreement then on technological advances.

I think here of the line from Jim Gill may all your problems be technical

p. 8-actually this is link to p. 7, since 8 is excluded

Another trend is the blurring of the distinction between primary sources, generally viewed as unprocessed or unanalysed data, and secondary sources that set data in context.

Good point about how this is a new thing. On the next page she discusses how we are now collpasing the traditional distinction between primary and secondary sources.

p. 6

Retrieval methods designed for small databases decline rapidly in effectiveness as collections grow...

This is an interesting point that is missed in the Distant reading controversies: its all very well to say that you prefer close reading, but close reading doesn't scale--or rather the methodologies used to decide what to close read were developed when big data didn't exist. How to you combine that when you can read everything. I.e. You close read Dickins because he's what survived the 19th C as being worth reading. But now, if we could recover everything from the 19th C how do you justify methodologically not looking more widely?

The "business" of univeristy?

les algorithmes ont besoin de données soi-disant neutres.. c'est un peu aller dans le sens des discours d'accompagnement de ces algorithmes et services de recommandation qui considèrent leurs données "naturelles", sans valeur intrasèque. (voir Bonenfant 2015)

p. 141

Initially, the digital humanities consisted of the curation and analysis of data that were born digital, and the digitisation and archiving projects that sought to render analogue texts and material objects into digital forms that could be organised and searched and be subjects to basic forms of overarching, automated or guided analysis, such as summary visualisations of content or connections between documents, people or places. Subsequently, its advocates have argued that the field has evolved to provide more sophisticated tools for handling, searching, linking, sharing and analysing data that seek to complement and augment existing humanities methods, and facilitate traditional forms of interpretation and theory building, rather than replacing traditional methods or providing an empiricist or positivistic approach to humanities scholarship.

summary of history of digital humanities

p. 100

Data are not useful in and of themselves. They only have utility if meaning and value can be extracted from them. In other words, it is what is done with data that is important, not simply that they are generated. The whole of science is based on realising meaning and value from data. Making sense of scaled small data and big data poses new challenges. In the case of scaled small data, the challenge is linking together varied datasets to gain new insights and opening up the data to new analytical approaches being used in big data. With respect to big data, the challenge is coping with its abundance and exhaustivity (including sizeable amounts of data with low utility and value), timeliness and dynamism, messiness and uncertainty, high relationality, semi-structured or unstructured nature, and the fact that much of big data is generated with no specific question in mind or is a by-product of another activity. Indeed, until recently, data analysis techniques have primarily been designed to extract insights from scarce, static, clean and poorly relational datasets, scientifically sampled and adhering to strict assumptions (such as independence, stationarity, and normality), and generated and alanysed with a specific question in mind.

Good discussion of the different approaches allowed/required by small v. big data.

p. 86

25% of data stored in digital form in 2000 (the rest analogue; 94% by 2007

Kitchin, Rob. 2014. The Data Revolution. Thousand Oaks, CA: SAGE Publications Ltd.

Kitchin, Rob. 2014. The Data Revolution. Thousand Oaks, CA: SAGE Publications Ltd.

are there non-digital data?

Really nice way to end the article.

I love Data = People :)

AMEN!

Interestingly enough, a lot of teachers either don’t know that such data might be available or perceive very little value in monitoring learners in such a way. But a lot of this can be negotiated with learners themselves.

Data=New Oil

Replication data for this study can be found in Harvard's Dataverse

So what data does h retain?

• username
• email address

Do annotations count as data?

Fair enough.

And "Burn!!" on Microsoft with that link.

Data in digital scholarship

This pithy statement brings together all sorts of previous annotations. Would be neat to map them.

A group of experts got together last year at Daghstuhl and wrote a white paper about this.

Basically the idea is that the data, the code, the protocol/analysis/method, and the narrative should all exist as equal objects on the appropriate platform. Code in a code repository like Github, Data in a data repo that understands data formats, like Mendeley Data (my company) and Figshare, protocols somewhere like protocols.io and the narrative which ties it all together still at the publisher. Discussion and review can take the form of comments, or even better, annotations just like I'm doing now.

Astronomy was area of collaboration because they needed to share data

This strengths the idea that data visualization is a field where a personal approach is still viable, as is shown also by a lot of individuals that are highly valuated as data visualizers.

nice bibliography!

Putting the Service-Profit Chain to Work

分清楚哪些是能知道的，哪些是不能知道的数据

It was an amazing presentation by Mr Cesar A Hidalgo, It was an eye opener for me in the area of data visualisation, As the national level organisation, we have huge data, but we never thought about data visualisation. You projects particularly pantheon and immersion is marvelous and I came to know that, you are using D3. It is a great job

Cough GigaScience Cough. See integrated GigaDB repo http://database.oxfordjournals.org/content/2014/bau018.full

Neat use of TAGS Explorer. Did do so, to interesting effects. Regular livetweet events may be more interesting. And there’s something to be said about NodeXL and COSMOS.

Great Principles of Computing<br> Peter J. Denning, Craig H. Martell

This is a book about the whole of computing—its algorithms, architectures, and designs.

Denning and Martell divide the great principles of computing into six categories: communication, computation, coordination, recollection, evaluation, and design.

"Programmers have the largest impact when they are designers; otherwise, they are just coders for someone else's design."

We should have control of the algorithms and data that guide our experiences online, and increasingly offline. Under our guidance, they can be powerful personal assistants.

Big business has been very militant about protecting their "intellectual property". Yet they regard every detail of our personal lives as theirs to collect and sell at whim. What a bunch of little darlings they are.

OER Data Report

That’s something we’ve been doing, yes.

Delete "preferably". Limiting the scope of text mining to exclude societal and commercial purposes limits the usefulness to enterprises (especially SMEs that cannot mine on their own) as well as to society. These limitations have ramifications in terms of limiting the research questions that researchers can and will pursue.

This statement is more generally applicable than just to TDM. Besides, "Encourage" is too weak a word here, and from a societal perspective, it would be far better if researchers were to retain their copyright (where it applies), but make their copyrightable works available under open licenses that allow publishers to publish the works, and others to use and reuse it.

The unusual data released strategy involving crowdsourcing on twitter, is discussed in more detail in this blog http://blogs.biomedcentral.com/gigablog/2011/08/03/notes-from-an-e-coli-tweenome-lessons-learned-from-our-first-data-doi/

See also this population genomics study in Nature Genetics that uses this data: http://www.nature.com/ng/journal/v45/n1/full/ng.2494.html See also this blog posting on data citation of this data (and related problems): http://blogs.biomedcentral.com/gigablog/2012/12/21/promoting-datacitation-in-nature/

The panda and polar bear datasets should have been included in the data section rather than hidden in the URLs section. Production removed the DOIs and used (now dead) URLs instead, but for the working links and insight see the following blog: http://blogs.biomedcentral.com/gigablog/2012/12/21/promoting-datacitation-in-nature/

More on the backstory and other papers using and citing this data before the Cell publication in ths blog posting: http://blogs.biomedcentral.com/gigablog/2014/05/14/the-latest-weapon-in-publishing-data-the-polar-bear/

As a novel and potentially controversial finding, the huge amounts of supporting data are depositedhere to assist others to follow on and reproduce the results. This won the BMC Open Data Prize, as the judges were impressed by the numerous extra steps taken by the authors in optimizing the openness and easy accessibility of this data, and were keen to emphasize that the value of open data for such breakthrough science lies not only in providing a resource, but also in conferring transparency to unexpected conclusions that others will naturally wish to challenge. You can see more in the blog posting and interview with the authors here: http://blogs.biomedcentral.com/gigablog/2013/10/02/open-data-for-the-win/

For more about GigaDB, see the paper in Database Journal: http://database.oxfordjournals.org/content/2014/bau018.full

ID: BPwPsyStructuralData SubjectGroup: BPwPsy Acquisition: Anatomical DOI: 10.18116/C6159Z

ID: BPwoPsyStructuralData SubjectGroup: BPwoPsy Acquisition: Anatomical DOI: 10.18116/C6159Z

ID: HCStructuralData SubjectGroup: HC Acquisition: Anatomical DOI: 10.18116/C6159Z

ID: SZStructuralData SubjectGroup: SZ Acquisition: Anatomical DOI: 10.18116/C6159Z

Sadly, there may not be much work on opening up data in Higher Education. For instance, there was only one panel at last year’s international Open Data Conference. https://www.youtube.com/watch?v=NUtQBC4SqTU

Looking at the interoperability of competency profiles, been wondering if it could be enhanced through use of Linked Open Data.

American Statistical Association statement on p-values

Books on data science and R programming by Roger D. Peng of Johns Hopkins.

Great explanation of 15 common probability distributions: Bernouli, Uniform, Binomial, Geometric, Negative Binomial, Exponential, Weibull, Hypergeometric, Poisson, Normal, Log Normal, Student's t, Chi-Squared, Gamma, Beta.

http://software-carpentry.org/lessons/<br> Basic programming skills for scientific researchers.<br> SQL, and Python, R, or MATLAB.

http://www.datacarpentry.org/lessons/<br> Managing and analyzing data.

This will be the case also for the data visualizations showed here, once the data is curated and verified properly. Still data visualizations can start a global conversation without having the full paper translated to English.

50 Years of Data Science, David Donoho<br> 2015, 41 pages

This paper reviews some ingredients of the current "Data Science moment", including recent commentary about data science in the popular media, and about how/whether Data Science is really different from Statistics.

The now-contemplated field of Data Science amounts to a superset of the fields of statistics and machine learning which adds some technology for 'scaling up' to 'big data'.

C'mon, civic technologists, government innovators, open data advocates: this can be a call to arms. Isn't the point of "open government" to bring people together to engage with their leaders, provide the facts, and allow more informed, engaged debate?

UT Austin SDS 348, Computational Biology and Bioinformatics. Course materials and links: R, regression modeling, ggplot2, principal component analysis, k-means clustering, logistic regression, Python, Biopython, regular expressions.

paradox of unanimity - Unanimous or nearly unanimous agreement doesn't always indicate the correct answer. If agreement is unlikely, it indicates a problem with the system.

Witnesses who only saw a suspect for a moment are not likely to be able to pick them out of a lineup accurately. If several witnesses all pick the same suspect, you should be suspicious that bias is at work. Perhaps these witnesses were cherry-picked, or they were somehow encouraged to choose a particular suspect.

Wow!

Python interface to the R programming language.<br> Use R functions and packages from Python.<br> https://pypi.python.org/pypi/rpy2

Guidelines for publishing GLAM data (galleries, libraries, archives, museums) on GitHub. It applies to publishing any kind of data anywhere.

• Document the schema of the data.
• Make the usage terms and conditions clear.
• Tell people how to report issues.<br> Or, tell them that they're on their own.
• Tell people whether you accept pull requests (user-contributed edits and additions), and how.
• Tell people how often the data will be updated, even if the answer is "sporadically" or "maybe never".

https://en.wikipedia.org/wiki/Open_Knowledge<br> http://openglam.org/faq/

Though it’s described in such a simple way, there might be hidden power in adding these tags, especially when we bring eBooks to the Semantic Web. Though books are the prime example of a “Web of Documents”, they can also contribute to the “Web of Data”, if we enable them. It might take long, but it could happen.

Useful examples. Could imagine something like Wikity, FedWiki, or other forms of content federation to work through this in a much-needed upgrade from the “Personal Home Pages” of the early Web. Do see some connections to Sandstorm and the new WordPress interface (which, despite being targeted at WordPress.com users, also works on self-hosted WordPress installs). Some of it could also be about the longstanding dream of “keeping our content” in social media. Yes, as in the reverse from Facebook. Multiple solutions exist to do exports and backups. But it can be so much more than that and it’s so much more important in educational contexts.

Thick data

And we need an ever-larger data set, right?

As usual, @AudreyWatters puts things in proper perspective.

Even those who adopt a taylorist model to education may understand that “it takes money to save money”.

Could see a lot of potential for Open Standards, including annotations. What’s not so clear is how they can manage to produce such ePub while maintaining their DRM-focused practice. Heard about LCP (Lightweight Content Protection). But have yet to get a fully-accessible ePub which is also DRMed in such a way.

Well-structured annotations can pave the way towards Linked Open Data.

L’interopérabilité est essentielle à l’avènement du Web des données liées (en éducation comme ailleurs).

data requirements

Among the most useful summaries I have found for Linked Data, generally, and in relationship to libraries, specifically. After first reading it, got to hear of the acronym LODLAM: “Linked Open Data for Libraries, Archives, and Museums”. Been finding uses for this tag, in no small part because it gets people to think about the connections between diverse knowledge-focused institutions, places where knowledge is constructed. Somewhat surprised academia, universities, colleges, institutes, or educational organisations like schools aren’t explicitly tied to those others. In fact, it’s quite remarkable that education tends to drive much development in #OpenData, as opposed to municipal or federal governments, for instance. But it’s still very interesting to think about Libraries and Museums as moving from a focus on (a Web of) documents to a focus on (a Web of) data.

The effectiveness of infographics, or any other form of communication, can be measured in terms of whether people:

• pay attention to it
• understand it
• remember it later

Titles are important. Ideally, the title should concisely state the main point you want people to grasp.

Recall of both labels and data can be improved by using redundancy -- text as well as images. For example:

• flags in addition to country names
• proportional bubbles in addition to numbers.

There are initiatives to do much more than this, including using Open Data on syllabi to delve down into course content.

https://github.com/rhiever/tpot TPOT (Tree-based Pipeline Optimization Tool) Built on numpy, scipy, pandas, scikit-learn, and deap.

All about open education

As long as it has some URI, it can be annotated. Any object in the world can be described through the Semantic Web. Especially with Linked Open Data.

First step in Linked Open Data is moving away from image PDFs.

Open Research MOOC

Not necessarily true. This is a misconception. "About 70 percent of OA journals charge no APCs at all. We’ve known this for a decade but it’s still widely overlooked by people who should know better." -Suber http://lj.libraryjournal.com/2015/09/opinion/not-dead-yet/an-interview-with-peter-suber-on-open-access-not-dead-yet/#_

In a nutshell, an ontology answers the question, “What things can we say exist in a domain, and how do we describe those things that relate to each other?”

According to inventor of the World Wide Web, Tim Berners-Lee, there are four key principles of Linked Data (Berners-Lee, 2006): Use URIs to denote things. Use HTTP URIs so that these things can be referred to and looked up (dereferenced) by people and user agents. Provide useful information about the thing when its URI is dereferenced, leveraging standards such as RDF, SPARQL. Include links to other related things (using their URIs) when publishing data on the web.

In section 4.1.3.2 of the xAPI specification, it states “Activity Providers SHOULD use a corresponding existing Verb whenever possible.”

Example of the type of harm that lack of transparency can lead to.

How can anyone argue that science isn't served by making primary data available? We must recognize that more people are harmed by not sharing data than are harmed by data being shared.

Allen Brain Institute