Some Questions Raised in Discussion (on day one)

We then divided into four groups and asked each group to bring their ideas together into a single blog entry for posting here. This brought together some of the thoughts from the lightning talks, the questions raised the day before and generally aimed at providing more concrete suggestions.

The result was the following four group-authored postings:

On Big Data (small data) and (trans)national sources

The Value of Visualisation and the Responsibilities of the Visualiser

Read Me First 

How to get through time, money, and institutional constraints

The group writing discussion exercise seems to have worked well, and was followed by reports from each of the groups on their ideas. The afternoon continued with tutorials on the use of QGIS by Hans Blomme and a presentation which highlighted the possibilities, performance, and challenges of using WebGL and Three.js with Tobias Englmeier.

1. How to reach efficiency in a given time constraint?

Most academic projects having strict deadlines (PhDs are more and more being given the limit of 36 months), deciding to include a wide DH dimension may bring substantial risks regarding the completion of the project on time. Building the database is a heavy process which does not even deliver a finished and visualisable product. Afterwards, analysing that database inevitably highlights the need to revise and rectify the database. This is the part of the project which might become time-consuming and hard to estimate. Therefore, throughout a new type of collaboration and teamwork (which is more hybrid and wasn’t possible 15 years ago) we have ways to minimise this.

2. How to convince institutions of the relevance of a DH project?

A research project is surrounded by institutions (supervisors, examiners, funding bodies, commentators, publishers…) who all have influences and expectations of results. They are actors who have approved your topic but have not necessarily allowed the emergence of a digital part – which cannot add up to a word count or allocated budget. For the researcher, the DH side is not a mere embellishment but a genuinely integral part of the academic output: this is the institutional constraint that has to be pushed back.

Once this is legitimised, DH workshops could be compared to academic writing workshops (where skills are acquired), and to a greater extent, data building would be seen as equally important as a literature review for instance. We could picture the appearance of ‘Digital Humanities’ in humanities curricula or in research guidelines.

1. Why do you want to make a visualisation?

Do you intend to make a specific argument with it? Do you intend to make a heuristic tool to explore a large collection of materials?

2. Is the data in your sources suitable for visualisation?

To begin with, is it in a digital format and thus not require digitisation before hand? In the case of map data is it geocoded? And just as importantly what is geocoded and is this consistent across the data? In the case of network visualisations are the relations between your object machine readable?

3. Are you considering variation across time, space, or both?

Are you focused on a particular state of affairs or the flow of a process? This will have repercussions on the design of the visualisation, including whether it is to be captured in a static or series of static visualisations or something dynamic, however that might be defined.

4. Are you prone to the sin of over classifying your data?

Can the design of your visualisation contribute to the naturalisation of problematic historical categories. The solution is not the proliferation of categories because this ironically increases the likelihood of including ones that are problematic.

To mitigate this problem: try to first identify more stable, and less problematic identifiers and first focus on them. At the very least add descriptive justifications for the structuring of the data in your visualisation and its limitations. It may also be productive to create a variety of visualisation based on the same data but which highlights some of the potentially different different perspectives.

4. If your visualisation will involve employing computative methods that will qualitatively transform the output and you do not understand the algorithms and the consequences of this transformation,

5. Keep in throughout the process that it is the researcher and not the visualisation which produces the analysis. Comment: There is some variation in the way this was to be expressed: knowledge vs. analysis, how much or how little to we concede these tools and visualisations can accomplish.

6. In the case of maps, be conscious of the way in which your projection, your perspective, your bounding limits, and other design choices dictate or at least delimit the way in which it can be read.

For example, be aware of how you map may reproduce certain relationships between center and periphery, etc.

7. Seek inspiration from other visualisations of different kinds that attempt to tackle similar problems or sources.

8. Be conscious of the time that may be required to gain the minimum competency in the tool you have chosen to create the map or visualisation. This is often underestimated.

You don’t have to have do everything yourself. The best visualisations are the product of close collaborations.

– Hans, Georgina, Alex, Gero, and Konrad

Our discussion focused particularly on a defense of the value of visualisation, bearing in mind the very legitimate questions and concerns regarding it which have been raised over the course of the conference, together with an awareness of the responsibilities which come with producing a valuable visualisation.  Without attempting to reduce our several, perhaps disjointed ideas to any too regulated form, we raised the following issues:

Defense of the Value of Visualisation

A manifesto for the small in Big Data and Digital Humanities – in questions (mainly)

Why do Data have to big Big?

What is the benefit of being big?

And when does big become big?

Where does big data start?

Can small data be beautiful?

How to get the best out of both worlds – big and small?

Who owns Big Data (intellectually)?

Is “cleaning” data an act of interpretation and hermeneutics?

Would it make sense to replace big by complex?

Is big data an elegant way of shying away from complexity?

Is big data flat – loosing edges, complexities

How to bring in our strenghts (historians, that is) trained as close readers into big data (kritische Haltung)?

Where is the competence of historians – same competences but bring them from analogue to digital?

You have the data – but we have the questions, ok?

Please let us in – or are we (historians) just the guys who annoy because we ask questions and seek Komplexitätssteigerung?

A problem of sources in transnational histories 

A high number of our sources are bound to national institutions (archives, national libraries). Could the maps and visualisations we have in mind allow us to de-centre or de-nationalise the nature of sources? A great example is the newspaper collection of the Austrian National Library in Vienna. However, as long as these collections remain bound to either a state (Habsburg mainly or a single language) we cannot de-centre or de-nationalise the objects of our studies. The visualisastion and mapping of events along such newspapers would only reinforce the existing dominance of national frameworks.

Anna Annieva, Bernhard Struck, Martin Stark, Stefan Nygard

• As historians, to what degree should we set aside time to acquire the skills and knowledge in some of the tools and methodologies that might enable us to engage with social network analysis and geographic analysis literature? How can we encourage the “hybrid” roles in projects involving mapping, visualisation, or large data components.
• When we create maps or visualisations in our work, what are the most effective means for tracking change over time in the objects/subjects we study?
• How do we encode ambiguity in our visualisations?
• What are some of the most effective ways to balance the quantitative and qualitative approaches as well as scales between micro and macro (the local and the global) in our projects?
• On the issue of scales again, how do we confront the challenges of making diverse sources, and diverse scales of source material comparable?
• What are some of the challenges and best practices in bridging and defining the disciplinary boundaries in historical work of a spatial nature or with strong computative or visualisation components?
• We have seen time and time again that assembling data and cleaning data are extremely labour intensive and highly underappreciated. How do gain recognition for this labour, and incorporate the costing and scheduling of it in the design of our projects and grant applications?
• How to we generate more open-ended research questions – with mapping, space and visualisation in mind?
• Lately we have seen PhDs and post-docs in Digital Humanities emerge (example here with KCL). Is that the way forward? Shouldn’t we let the ‘DH’ label fade away and make sure everyone is “DH-ready”?
• What are maps and visualisations? Are they part of our “narrative” and thus analysis? Are they an “archive”?
• Do maps and visualisations replace narrative or rather enhance our written narrative? What is the relation between text and image in our histories?

QGIS Downloads

For Mac OS X users, please download and run the installer for the GDAL Complete, the Matplotlib Python module (other python modules there are optional), and QGIS, which can all be downloaded here:

KyngChaos QGIS Installs.

Hans has uploaded some exercises and material for us to download before the workshop:

Workshop Files 1

Workshop Files 2

1. Download and install a local web server on your laptop, for example XAMPP

2. Go to the following link and download the zip-folder to your harddrive.

3. Extract the zip-folder into your webserver’s root directory (on XAMPP this is called “htdocs”) so that you have a file structure like …\htdocs\Three_js_layer_workshop\index.html

4. Start your webserver (On XAMPP open the control panel and click “Start” in the first row besides where it says Apache).

5. When the server is running, you can access the site by entering the following url into your browser: localhost/Three_js_Layer_Workshop

If everything works, you should see an empty black map with its center over Scotland.

The idea of the workshop is to create a sample visualization with three.js Layer, which should make visible the  a dataset from the addressing history api. The dataset (comes with the zip-folder) acommodates 10.000 geo referenced historic adresses from the Scottish Post Office Directories.

For Mac OS X users, you can find more information about this here:

OS X Java Control Panel

Windows users please see:

Windows Java Control Panel

Some other instructions related to the installation of Java:

https://www.java.com/en/download/help/mac_java_update.xml

https://www.java.com/en/download/help/index_installing.xml

https://www.java.com/en/download/help/download_options.xml

For the installation of VennMaker itself please see:

Download VennMaker 1.4.0 Trial Version

The overall aim of our contribution is to evaluate the digital checks and balances of collaborative research in the study of transnational connections in the field of social reform in the long 19^th century.

Between 1815 and 1914 new trans-organizational milieus consisting of journals and conferences became important carriers of ideas and contention. They can be seen as laboratories for the development and diffusion of reformist ideas in Europe.Progressive liberal in origin, conferences for instance were used a part of a repertoire to contest existing power relations and the social and cultural status quo. Our interest ties in with the turn towards a relational and transnational approach in the study of the ideas, actors and networks that shaped welfare state politics and policies in Europe and beyond. Transnational history, however, not only takes place in transnational spaces. We are, therefore, particularly interested in the relations and interactions between the spaces where transnational actors are active and the identification of so called ‘rooted cosmopolitans’, described by Sidney Tarrow and della Porta as “people and groups who are rooted in specific national contexts, but who engage in regular activities that require their involvement in transnational networks of contacts and conflicts”.

We see that while researchers of social reform movements and collective action are increasingly adopting network concepts and perspectives in their work, their use of formal network analytical tools for the exploration how networks of people and discursive frameworks come into existence, is still very limited. This is due to heuristic barriers, such as the identification of comparable sources of data across borders. We are interested in contexts and conditions of knowledge transfer, which we study in networks experts and activists, but also in the knowledge itself: from the theoretical foundations to views and practical ideas. For this purpose it should be possible to mine our sources for changing concepts and discourses. Therefor we have recently founded the Virtual Research Environment (VRE) ‘TIC-Collaborative’ (see www.tic.ugent.be). TIC Collaborative is a Virtual Research Environment for the study of 19^th and early 20^th century international organizations and congresses. The VRE applies the principle of scholarly crowdsourcing (as already used by eight research units across Europe and 4 PhD students) and offers access to digitized dispersed sources in a central observatory. The platform includes a (1) digital asset management system which will allow the research community to share, annotate and enrich a wide variety of documents, letters, photographs, etc. and (2) a collaborative relational database (in order to process the data for social network analysis, prosopography, etc.). It thus integrates digitized primary sources concerning international congresses (e.g. conference proceedings, reports, lectures and attendees lists), international organizations, and other published and unpublished sources (e.g. yearbooks, periodicals, articles of association, pamphlets, and memoirs).

1. Collaborative digital asset management system. Most digital humanities projects tend to start out from existing collections, thus limiting the degree to which they can operationalize digital techniques to the specific questions that should interest historians. Instead, in our approach the corpus selection is research driven. Therefore we’ve been compiling a corpus of dispersed sources on long 19^th century social and legal reform. Our intentions go beyond just compiling the sources, rather we are building a platform that enables the researchers to combine the use of both ‘traditional’ qualitative approaches as well as digital methods. Many of the advanced mining techniques—such as forensic analysis of texts (author recognition), topic modeling, text clustering, named entity recognition—are currently still in an experimental stage, but very recent findings show the exciting potential of for example multi-lingual text-mining in large historical datasets and text corpora (Huijnen et al. 2014). Exploratory searching can provoke new questions and research objectives, but in fact until now it has mainly been used for the confirmation of existing historical knowledge (Gibbs and Owens 2012). We aim to avoid this by using a hybrid approach, combining text mining (topic modeling in particular) with network analysis, data visualization techniquesBörner and Polley 2014) and ‘traditional’ close reading (Verbruggen and Carlier 2014). TIC-collaborative allows the exportation of enriched data in order to perform historical discourse and linguistic analysis with other tools such as NVivo (software for qualitative data analysis) or the Stanford Topic Modeling Toolbox. Cited concepts, persons, publications, and educational experiments are analysed with Sci2 (https://sci2.cns.iu.edu/user/index.php), a toolset specifically designed for spatio-temporal visualization of knowledge circulation.

(2) Another feature of our VRE is a joint relational database, which links international congresses, international organizations, people and publications in order to process the data for social network analysis, cluster analysis, prosopography. During the last several years, various initiatives were initiated to build a structured database about persons, organisations and congresses in the field of social and legal reform, such as at the prosopographical database of magistrates (UCL), TIC (Ghent and Maastricht University), ODIS (www.odis.be; KUL, Ghent and Antwerp University, VUB) and the Europhil group (EHESS, Maastricht University,…). It is our goal to integrate these data in one single platform. We also envisage the exchange/linking of data with other databases and a wide range of query- and export possibilities (Json, XML, xls). This part of the VRE is already operational and uses the ‘Nodegoat’ software. Nodegoat is a web-based database management, analysis and visualisation platform: “Using this system, scholars define, create, query, update, and manage any number of data tables by use of a graphic user interface. Within nodegoat, scholars are able to instantly analyse and visualize data sets (also in space). It allows scholars to enrich data with relational, geographical and temporal attributes. Therefore, the modes of analysis are inherently diachronic and ready-to-use for interactive maps and extensive trailblazing.” (www.nodegoat.net). Next to these built-in functionalities, Nodegoat also enables the researcher to export datasets, ready to use in more advanced visualization software such as Gephi and UCINET (network visualization) or QGIS and ArcGIS (GIS).

Beyond the history of social reform and the attempt to bring particular collections and databases together, the use of a VRE allows the development and assessment of methodologies in the field of digital humanities. Therefor the TIC-Collaborative VRE will serve as a pilot case within the recently founded Flemish/Belgian branch of DARIAH (Digital Research Infrastructure for the Arts and Humanities) (DARIAH-VL/BE). Researchers in the humanities who want to make use of digital analytical tools often face two problems: the tools are developed in an ad hoc manner within the scope of a particular research project and are not evaluated or communicated outside the project context. Secondly, research projects require specific types of scholarly data resources in combination with specific tools. To match tools and research questions is therefore a complex activity. In order to valorize analytical tools better and exchange expertise, DARIAH-VL/BE aims to set up a common standards-based infrastructure offering these tools as a service for all research projects, together with a documentation knowledge base of tools, standards and best practices. In this way, previously developed tools can be reused and may even become ‘standard tools’ in the field. Being a pilot project within DARIAH-VL, TIC-Collaborative will contribute to this.

Download as Word File:

TIC Collaborative – project (2_6_2014) TIC-Collaborative, Transnational Intellectual Cooperation in the long 19th century

Further Material:

TIC Collaborative – project (2_6_2014)

Map for Discussion: