but increasingly primarily machines,

There are ongoing and fierce debates on what exactly constitutes a 'persistent' identifier

Through a well-defined protocol

Data should be Findable Data should be Accessible Data should be Interoperable Data should be Re-usable.

In reactions to the Turing test, one may easily discern a fear of machine intelligence underlying many of the counterarguments.

But even as these arguments are advanced, the detractions seem obvious

Grants in excess of fifty thousand dollars are quite rare in the humanities; grants exceeding twice that amount are by no means unusual in DH

Increasingly, people who publish things online that look like articles and are subjected to the usual system of peer review need not fear reprisal from a hostile review committee

articles about “new media” seldom

Latino precincts are those in which at least 50% of the registered voters are Latinos. The actualLatino population is higher, especially in areas with high concentrations of Latinos without UScitizenship

including number of registered Democrats and Republicans,and ethnic breakdowns of registered voters and the voting age population. Contextual datafrom the 2000 census by census tract were linked with the precinct-level data.

a consequence of the post-Cold War restructuring of the aerospace industry in Los An-geles

confusion between residence in the City or County of Los Angeles may reduce ease of politicalparticipation

applied automatically over a large corpus of digitized texts.

we start by finding references to specific political divisions

can locate place names using the document's context.

when a text mentions Paris, does the writer mean Paris, Texas, USA or Paris, France

Such visualizations are possible because of the datasets that have been gathered by scholars over many decades

But both methods share a similar vision and a similar blindness.

aimed to bring the hidden patterns in the data on lynching to light.

a focus on the designer’s ownsubject position can help to expose the decision

Is our data the right type?

One of the key contributions of STS has been to challenge the ideathat science and/or technology is objective and neutral by demonstrat-ing how scientific thought is situated in particular cultural, historical,economic, and social systems [77

People are linked through the groups they belong to. Groups are linked through the people they share.

links between people and some other kind of thing

All I know is whether someone was a member of an organization or not.

has been used to collect and analyze information on more than two hundred and sixty persons (of varying degrees of suspicion) belonging variously to seven different organizations in the Boston area.

none of the information that I’ve just told you is immediately evident in the letter on the screen.

For instance, here’s an 1801 letter from Jefferson to William Evans, an innkeeper friend. You can see in this letter that Jefferson makes reference to a “former servant James,” underlined in green. If you click the link, it takes you to an editorial note that informs us that this “former servant” is in fact James Hemings, who, as it turns out, three years after the agreement with Jefferson was recorded and witnessed, finally received his freedom

thinks that Hemings most likely made a verbal agreement with Jefferson to return to America

How does one identify and extract meaning from the unique set of documents that do remain

this data to pursue their own lines of inquiry. Getting a broad sense of where the field has b

oreover, the overall increase in work on women’s history does not translate into an equitable

ting schema (such as Library of Congress subject headings), topic model

pic modeling, a computer science data mining technology that is arguably the state-of-the-art model for text document collections, allows for a m

. America: History and Life (hereafter AHL) focuses on the history of the geographic regions th

By highlighting the process of topic modeling, Goldstone and Underwood reveal how different methodological choices may lead to contrasting results.

Rhody’s work is perhaps the best evidence thus-far that what we might have identified as cohesive “topics” are more complex than simple thematic connections

It describes a method of extracting clusters of words from sets of documents

before realizing there isn’t much immediately apparent you can actually do with it

You’re introduced to topics, and how a computer came to generate them automatically without any prior knowledge of word definitions or grammar.

treating the works themselves as unceremonious “buckets of words,” and providing seductive but obscure results in the forms of easily interpreted (and manipulated) “topics.”

we conclude the chapter by saying that all the sentiment analysis tasks are very challenging

The existing research assumes that the document is known to be opinionated

In many cases, opinions are hidden in long forum posts and blogs. It is difficult for a human reader to find relevant sources, extract related sentences with opinions, read them, summarize them, and organize them into usable forms.

One of the main reasons for the lack of study on opinions is the fact that there was little opinionated text available before the World Wide Web

In this chapter, we only focus on opinion expressions that convey people’s positive or negative sentiments

Most tools at our disposal have weak or nonexistent semanticcapabilities; they count, compare, track, and represent words, but they do not produce meaning

Computers can help extend human reading and understanding, especially for largecollections of texts that you couldn't read in a lifetime

With such rich and sophisticated analytic environments, do we even need to read texts anymore?

They provide a snapshot,but do not allow exploration and experimentation

As a result,there is a significant inequality in the availability of digital texts, one that has a profound effect on the kindsof work that scholars are able to pursue

struggling with how to derive meaning from texts, fromhigh-school students researching an essay topic to journalists combing through leaked security documents,or from companies measuring social media reaction to a product launch to historians studying diversity ofimmigration based on more than two centuries of trial proceedings

were it not for text-based searches of the title, description, andother metadata.

You type in a word to search for and the interactive returns a simple bar graph that youcan drop into a comment

One of the stylistic breaks takes place in the 1870s (i.e. after the Civil War), the other in the 1920s (in the period of prosperity before the Great Depression); the third peak is not fully formed yet, even if one can observe an acceleration of language change at the end of the 20th century

Then we dismiss the original hypothesis, in order to test new ones: we iterate over the timeline, testing the years 1836, 1837, 1838, 1839, … for their discriminating power

The procedure randomly picks n text samples written before and after the assumed break;

Since none of the out-of-the-box techniques is suitable to analyze temporal datasets

is a tacit assumption that the researcher knows in advance which elements of the language are subject to change.

Although some of these instances might refer to a body of flowing water, there is no guarantee that they all use the word in the same sense

statistical methods in applications like search engines, spellcheckers, autocomplete features, and computer vision systems.

I argue that, when applied to the study of literary and cultural texts, statistical text-mining methods tend to reinforce conceptions of language and meaning that are, at best, overly dependent on the “literal” definitions of words

The odds of reporting a high level of psychological strain increase with age

It appears that work has a different meaning for different people.

The data are taken from the 1992 British Household Panel Study (BHPS), which includes data on 7897 individuals.

level of general happiness.

individual characteristics on measurements of psychological well-being

that food prices were positively correlated with workhousedeath rates,

Indeed, no measure isentirely immune to outside conditions.

Thequestion is inherently relative

A workhouse in whicheveryone died shortly after admittance might seem badly managed, butthat would not necessarily be the case if it attracted only those in themost extreme state of need.

Thus the population in a workhouse is a‘choice-based’sample

Though at its peak 3 million people out ofa total population of 8.5 million were dependent on relief, for people incomfortable circumstances life went on more or less as normal.

thereareways ofaddressing this most likely problem thatoftensucceed

variationbetweenpeople, not variationwithin individuals over time.

over time.

Globally, there appears tobe bias toward males, but when individual graduate schools aretaken into account, there seems to be bias toward females.

A treatmentthat appears effective at the population-level may, in fact, haveadverse consequences within each of the population’s subgroups

We show that Simpson’s paradox is most likely to occur wheninferences are drawn across different levels of explanation (e.g., from populations tosubgroups, or subgroups to individuals).

he sample size of Census data isalso considerably larger than the sample size of GSS data

researchers can classify respondents as same-sex behaving only if theyare in a cohabiting relationship,

public policymechanisms must increase the number of tolerant workplaces and not simply punishdiscriminating firms

Wage differentialsfor same-sex behaving men are surprising because sexual orientation is not a visible trait.

These states represent 68 percent of released prisoners in 2005, 69 percentin 2010, and 67 percent in 2012

Only 23 states provided data for the entire 2005-15 time frame

collects data submitted voluntarily by state departments of corrections and parole

l and legal environment in which they must operate. To deal with such issues, it

Another finding is that death did not choose people at random. Analysis of data for Trinidad indicates that the annual death rate for the shortest quintile of

from 1750 to the present. One of the principal findings to date is that native-born Americans reached modern levels of height and nutrition by the time of the American Revolution, but there were long periods of declining nutrition and height during the

Tbis is Tr.uitor tlui;:e Ct:nrntit:s from now

Add to lhis the:.-known probabiJity of I mpt:ria.1 assassi-nation. viceregal tevolt. the conrem1>0rary recurrence of pei·iods <>f economjc depression, the dc.:cli.oing: f;.tte of pboct:u-y c....xplo• rations, the:

you wiJJ le:lrn ro ap-pJy psychohisrory to all probk111s ;i,.; :1 11:uler of course

data gathered by the corporation is not nearly as complete or accurate as that gathered by community organizations in the state

chose to purchase California eviction data covering the same areas for $100,000

but we never received an adequate response.

Some of the groups

we have encountered major problems with Eviction Lab’s practices of big data production

it only cares that you have an identity that is addressable by Facebook.

individual behavior is regulated through the enforcement of norms

they were still engaging in an act of identification

The morecomplex the data structure of a computer program, the simpler thealgorithm needs to be, and vice versa.

Computeralgorithm games, for instance, are experienced by their players as narratives.

Computeralgorithm games, for instance, are experienced by their players as narratives.

t is more interesting to think about the ways in which search engine results perpetuate particular narratives that reflect historically uneven distributions of power in society

Published text on the web can have a plethora of meanings, so in my analysis of all of these results,

In the case of Google's history of racist bias

Understand which domains are goto when it comes to finding domain experts.