>> From the Library of Congress in Washington D.C.
>> Our final talk before the break is from Stephen Robertson
who is Director of the Roy Rosenzweig Center for History
and New Media and Professor History at the George Mason University.
He will tell a story about Digital Harlem,
a project that won the American Historical Association's Roy
Rosenzweig Prize for Innovation in Digital History
and the ABC CLIO Online History Award
of the American Library Association in 2010.
You're on, Stephen.
>> Stephen Robertson: Good morning.
[ Applause ]
Applause [inaudible].
I just want to start with a quick shout out to the wonderful team
of people here at the Library of Congress
who have brought us all here today, and a fascinating group
of people they've brought.
I'm going to go back to maps which it shows you a lot at the beginning
to talk about data in place.
Historical sources are full of data about places and locations.
The spatial data is not very intelligible in textual form,
and even when extracted and organized in tabular form,
it really doesn't tell us a lot.
It's hard to discern the story that it has to tell us.
Now, mapping has been around for a long time as a tool for making sense
of the spatial data, but it's only really with the advent
of web mapping that it's become accessible
to a wider range of people.
You used to have to go find the GIS person in your university
to make you a map for a project, and that was going to be the one map
that you did if you had the months to do that.
The web has fundamentally transformed that, and the project
that I'm here to talk about today owes its form
to the launch of Google Maps in 2005.
Shane White put together a team of four historians,
University of Sydney, of which I was the junior member,
to do a study of everyday life
in New York City's Harlem neighborhood in the 1920s.
Our key sources for this to get beyond the political
and artistic elite that dominate accounts
of Harlem was New York City's two black newspapers, The Age
and the Amsterdam News and the records
of the Manhattan District Attorney.
Now, I was in that project because I'd used those legal records
in my dissertation, and I knew from the dissertation that they were full
of information about where things happened that I'd found no way
to use and analyze when I did the dissertation.
So, what I suggested as my contribution
to this larger project was an effort to map our sources.
Now, when we got funding for the project in 2004, the technology
for doing that was our GIS, and I went away to find collaborators
at the University of Sydney,
our experts in art GIS, to make that happen.
And, at Sydney at that time,
it was the archeological computing laboratory.
However, the problem with art GIS in 2004 was
that it was not possible to use it on the web.
It wouldn't run on our beloved Macintosh computers,
and it simply was, as it still is, far too hard to master
and far too complex to make it worth using
to analyze the qualitative data that we had.
So, and, thankfully, with the launch of Google Maps in 2005 and thanks
to Damian Evans, one of our team who created a hack between our database
and Google Maps, we launched Digital Harlem in 2009
as a web based form of mapping.
As with any large digital project,
there's a whole team of people behind that.
And, before I say anything more, I want to acknowledge all
of those people in addition to the ones I've already named,
a team of historians, graduate research assistants,
technologists of various kinds, and almost a million dollars
of Australian government money which always amuses people.
But, the Australian government, at least in the early 2000s,
was interested in funding innovative scholarship even
about black neighborhoods thousands of miles away.
Now, Digital Harlem was one
of the first historical we mapping projects,
and one of the first digital history projects to shift
from what the Valley of the Shadow was really interested in doing,
digitizing material, creating online collections towards visualizing
those sources.
Now better technologies now exist for web mapping projects.
Nevertheless, Digital Harlem remains a useful starting point for thinking
about mapping as a means of making sources more accessible,
more visual, and more useful.
The contrast here is research with access
and sources through a database.
I think Michell Whitelaw's formulation which I'm sure many
of you are familiar with, captures this best.
Search is ungenerous.
Demanding a query and returning only the terms you enter.
Withholding information about the structure and materials available
and filtering out any alternative hypotheses.
Visualizations, by contrast, are generous, rich, browsable interfaces
that reveal the scale and complexity of the data behind them
and provide a context, a context that enriches the exploration
and analysis of that data.
So, Digital Harlem is not an interface to a collection of forces.
Thanks to copyright, licensing, and restrictions imposed by archives,
it's not possible for us, even if we'd wanted to, which we didn't,
to offer access to the sources on which the site draws.
But, those restrictions do not prevent the creation of data
from those sources which is what Digital Harlem contains,
information about events, an ill-fated tennis tournament
in this case, about people, and to a much lesser extent, about, sorry,
much lesser extent about people.
A lot about places.
It's worth noting that those of us who work on the 20th century
and beyond encounter this problem with access
to sources far more often than our colleagues
who work in earlier periods.
We need to be talking more about the way in which restrictions on access
to sources shape the kind of digital projects we can do
on the 20th century and how creating data is a way of getting access
to material that's otherwise under copyright.
To create data, however, requires a different engagement with sources
than humanity scholars typically have.
Whereas, to use Miriam Posner's words,
we usually immerse ourselves in our sources.
Dive in. Understand them from within.
To create data is to extract information.
And, features from sources requiring a decomposition of a subject
or object into attributes and variables.
If you're now increasing range of computational tools to do
that extraction for you, the data that we gathered
for Digital Harlem was done by hand.
And, because of the problems of access to digitized newspapers
and the limits of ACR, it's still a process
that needs to be done by hand.
So, myself and the team of research assistants who appear briefly
on the screen recorded details of every location
and every event associated with the location in those legal records
that I was talking about.
And, in the Harlem's two black newspapers.
And, crucially, in a way that Ed was alluding to, not just information
from the news stories, but information
from every section of the newspapers.
The fraternal reports, the church records, the sports pages.
There's far more information and far more spatial data in newspapers
than we're used to thinking about when we roll the microfilm through
and looked for the news stories that were part of that.
We organized that information into a data model,
entered it into a database, and geocoded it so that could be mapped.
Now, the metaphor that we commonly use to describe that process,
mining, sits awkwardly with humanity scholars concerned with development
of empathy with an appreciation of the position of the person or group
or the qualities of an object.
It sets up a sense of creating data as somehow dehumanizing.
Mapping data can somehow mitigate that consequence,
can align with an orientation towards [inaudible] data,
one element of which is articulated by [inaudible] as an appreciation
of context, interdependence, and vulnerability.
Maps as a visualization express this relationship of parts one to another
and to many to a greater whole.
Mapping data, mapped data is seen at its geographical context,
and [inaudible] Digital Harlem by the use of a historical map player.
A Bromley Real Estate atlas which shares a lot of the information
and qualities of the [inaudible] and Sanborn fire insurance maps.
It provides building footprints, information on the height
of those buildings, the material they're made out of,
whether they have shops or stores in them.
What it does most dramatically in terms of the maps that we're used
to looking at in historical scholarship is filled
in the spaces on a street map.
And, in that way, the space literally divides,
helps to subdivide and divide Harlem into multiple smaller places
and to give some indication of how those places interact.
And, if you know anything about the history of Harlem in the 1920s,
I chose this flock because on the corner is renaissance ballroom
and casino, the sight of a lot of dances and entertainment
and basketball games next
to the establishment Abyssinian Baptist Church,
one of the main line middle class moral racial [inaudible] of Harlem.
And, the hall to the right of that was the former headquarters
of Marcus Garvey's black nationalist UNIA.
A very interesting gathering of people walking down those blocks
in a way that we don't appreciate without this kind of level of data.
One of the very exciting things going on right here at the Library
of Congress is the digitization of Sanborn fire insurance maps.
We're just going to put these incredibly rich sources
into people's hand, make them freely accessible, transform the kind
of historical mapping projects we can do
by adding this incredibly rich layer of data
about what the place is like.
Layer the different data, and hence large quantities
of data can be combined on a single map, providing an image
of the complexity of the past.
And, I'm going to come back to this really complex looking map.
You can examine maps of sources at different scales, make comparisons,
discover relationships by visually detecting spatial patents
that remain hidden in those texts and tables that we started with.
Now, mapping data has become an increasingly common form
of interface for a lot of digital collections.
But, too often, those projects map only a single collection
of material.
That approach takes really only a small part of the power of mapping.
Geographic location provides a means to integrate material
from a wide range of disparate sources.
So, what's important about assigning a geographic reference to data,
as Karen Kemp puts it, then becomes possible to compare
that characteristic of the phenomenon, etc. with others
that exist or have existed in the same geographic space.
What were previously unrelated facts become integrated and correlated.
The power of maps to bring disparate things together is
where we really need to be going with them as a technology.
Used in this way, the geospatial web can help us capture the confluence
of multiple rhythms that [inaudible] argues make up everyday life.
So, this is one of my go to examples for talking
about Harlem in the 1920s.
It's a map of nightlife during prohibition using data
from newspapers, undercover investigations, legal records.
It shows the venues which drew crowds to Harlem, the night clubs,
the speakeasys, and the venues that black residents opened
in their apartments, known as buffet flats,
catering exclusively to blacks.
The map highlights the different geographies of those venues and,
in particular, the clustering of buffet flats in sections away
from the other venues, away from whites, and drew our attention
to how blacks developed spaces apart from whites as they flocked
to Harlem's night life in the 1920s.
Now, while this map captures some of how prohibition shaped night life
and leisure in Harlem, it's only a partial map
of the commercialized leisure available in the neighborhood.
Digital Harlem lets you create the context by adding dance halls,
theaters, pool rooms, the halls that hosted basketball games
and boxing bouts, and then you can add
to those commercial venues [inaudible]
to understand night life, all of the places
where noncommercial leisure took place in Harlem in this period.
Meetings of church groups, fraternal lodges,
community organizations, and social clubs.
You end up with this incredibly complex map.
Social clubs.
Incredibly complex map
which highlights fundamentally just how a small segment of leisure
and night life in Harlem actually appears in discussions focused
on prohibition which defines the way
that we understand what Harlem was like in the 1920s.
In regards to mapping events rather than places,
this is one of the first maps we created, arrests for numbers
in 1925, which gave our research a new focus.
It shows in the first instance the sheer pervasiveness
of numbers gambling which is a picture that a multitude
of sources reinforced in Harlem, and it's one reason why Shane White,
Steven Garton, and Graham White, and I wrote a book
about the wide ranging economic and cultural role of numbers gambling
in Harlem as one of the outcomes of our collaboration.
Zooming in to that map highlights how placing bets was woven
into everyday life.
Arrests occurred on street corners as residents get on their way
to work, and the businesses lining the avenues
as they went shopping and ran errands.
And, in their homes, on the cross streets
as numbers runners went door to door collecting bets.
Our original concept for Digital Harlem also included mapping
individual lives, but assembling enough data to reveal more
than a single moment
of an individual's life proved beyond our resources.
However, we did generate maps for a handful of people based
on information on their residence, work, and leisure
and probation and parole records.
Those maps make visible the extent to which the lives
of Harlem's residents were not bounded by the neighborhood,
made clear that the census data that we commonly rely on to determine
where people lived ultimately only tells us where they slept.
So, for example, during the five years that Morgan Thompson was
on probation, his work as a laborer took him not only outside Manhattan
but to the Bronx, Queens, and Brooklyn.
We've used lines linking residents with the workplaces and other places
that people frequented while [inaudible]
to suggest their movement through the city.
The geography of work was often different for women.
Annie Dillard, like the majority of Harlem women in the workforce,
found paid employment as a domestic servant in homes
on the upper west side, hotels in midtown,
and in a laundry in lower Manhattan.
The newest version of Digital Harlem adds a timeline to this
to understand how these lives evolved.
Two strikingly different geographies,
two strikingly different reminders that living
in a city is not living in the neighborhood.
It's moving across the city in a way
that we don't often place African Americans in places like New York.
Now, for all the maps that we feature on the site
and that we discuss in our scholarship, much of the usefulness
and impact of Digital Harlem comes from how it allows users
to make their own maps, to visualize the data that interests them rather
than being constrained to what interests the site's creators.
That capacity highlights that these maps are exploratory,
not illustrative, that they raise questions rather
than answering them.
But, the site itself, unfortunately, offers limited help in making sense
of a map and the data it visualizes.
That design is in keeping with the original conception of the site
which was as a research tool for those
of us collaborating on the original project.
When we decided to share the site, we added some material
about the places and events featured on the site
and on the individuals whose lives can be mapped.
And, we created a blog that linked to the site with posts
about additional maps of places and events such as traffic accidents
in this example that incorporate additional material
like photographs.
Unfortunately, those after the fact efforts can only go so far.
Creating a real context for understanding the data
in Digital Harlem is a project in its own right
that will require a wholesale redesign of the site,
and it's one of the things that we need to think about the difference
because producing sites that are research tools for people
who understand the data at some level and sites that are shared
and are going to be used by people
who don't bring the researcher's understanding to the site.
Now, an emerging offer, option for making data more accessible
and useful is to use what's in Digital Harlem as the basis
for a spatial narrative rather than simply a map.
I'm currently exploring that approach for a project
on the 1935 Harlem Riot for which we're creating another version
of Digital Harlem based on that single year.
By the 1930s, there's a much greater wealth of information about life
in Harlem than there is in the 1920s.
Now, a platform like the widely used story map [inaudible]
which creates a linear single path
through a map cannot effectively tell the story of a complex event
like the riot in which multiple things happened at any given time.
Neither, for that matter, can following the timeline
on Digital Harlem let you understand what is going on in the riot.
Neatline, a mapping plugging for [inaudible] which Ed was talking
about earlier, gives some scope for more complex storytelling,
and I've created a prototype narrative of the riot just
to think about, there we go, just to think about what's possible.
What I really like about Neatline is that the timeline slider
at the bottom of that image provides a means of navigating the exhibit.
Dragging it not only changes the points that appear on the map.
It also alters the way points visible on the right.
While points display information on particular events, way points,
way points can be used for broader arguments
which can be grouped together rather than being tied
to a single point on the timeline.
What that means is that it gives you some flexibility
in how a narrative is reared.
You can roll over each way point in a group or explore them
in a sequence, or out of sequence.
Each way point can be associated
with a zoom level on a specific location.
Clicking on a series of way points can, thus, move you around a map.
And, you can annotate those way points, attach them to polygons
and lines as I have in this prototype to draw attention
to the analysis of space in my understanding of what's going
on in the riot, to movement, direction, proximity,
connection, and patterns.
Used in that way, annotations shift some of the argument
into visual form, and that's ultimately where I think we're going
with the kinds of visualizations we've been making in maps,
that the future direction in visualizing data using maps
that this prototype points to is the capacity more extensively
and dynamically integrate maps in narratives.
To visually combine data
and interpretation while retaining the orientation towards putting data
in context, which, to me,
is the most powerful thing that mapping lets us do.
Thank you.
[ Applause ]
>> This has been a presentation of the Library of Congress.
Visit us at loc.gov.
For more infomation >> 3 Nasihat - Digital Storytelling 2017 - Duration: 6:11. 
For more infomation >> Digital Access Demo - Duration: 0:56. 
For more infomation >> Digital Marketing Introduction - Learn Online Course for free - Digital Marketing Steps - Duration: 3:27. 
For more infomation >> Hackathons - Digital Campus Bordeaux - Duration: 1:40. 
Không có nhận xét nào:
Đăng nhận xét