An Ecosystem for Linked Humanities Data
Rinke Hoekstra
Vrije Universiteit Amsterdam/University of Amsterdam
rinke.hoekstra@vu.nl
Albert Meroño-Peñuela, Kathrin Dentler, Auke Rijpma, Richard Zijdeman and Ivo Zandhuis
legenddatalegenddata

The Promise of Digital Humanities

The Promise of Digital Humanities

http://schoolofherring.com

http://science-all.com/fishing.html

The Problem of Digital Humanities
Pacific Barreleye, http://imgur.com/gallery/Mzyb5
(can rotate its eyes forwards or upwards to look through the transparent head to prey above)

http://www.asergeev.com/pictures/archives/compress/2012/1034/24.htm

The Cost of Data Preparation
Common Motifs in Scientific Workflows:
An Empirical Analysis
Daniel Garijo⇤, Pinar Alper †, Khalid Belhajjame†, Oscar Corcho⇤, Yolanda Gil‡, Carole Goble†
⇤Ontology Engineering Group, Universidad Polit´ecnica de Madrid. {dgarijo, ocorcho}@fi.upm.es
†School of Computer Science, University of Manchester. {alperp, khalidb, carole.goble}@cs.manchester.ac.uk
‡Information Sciences Institute, Department of Computer Science, University of Southern California. gil@isi.edu
Abstract—While workflow technology has gained momentum
in the last decade as a means for specifying and enacting compu-
tational experiments in modern science, reusing and repurposing
existing workflows to build new scientific experiments is still a
daunting task. This is partly due to the difficulty that scientists
experience when attempting to understand existing workflows,
which contain several data preparation and adaptation steps in
addition to the scientifically significant analysis steps. One way
to tackle the understandability problem is through providing
abstractions that give a high-level view of activities undertaken
within workflows. As a first step towards abstractions, we report
in this paper on the results of a manual analysis performed over
a set of real-world scientific workflows from Taverna and Wings
systems. Our analysis has resulted in a set of scientific workflow
motifs that outline i) the kinds of data intensive activities that are
observed in workflows (data oriented motifs), and ii) the different
manners in which activities are implemented within workflows
(workflow oriented motifs). These motifs can be useful to inform
workflow designers on the good and bad practices for workflow
development, to inform the design of automated tools for the
generation of workflow abstractions, etc.
I. INTRODUCTION
Scientific workflows have been increasingly used in the last
decade as an instrument for data intensive scientific analysis.
In these settings, workflows serve a dual function: first as
detailed documentation of the method (i. e. the input sources
and processing steps taken for the derivation of a certain
data item) and second as re-usable, executable artifacts for
data-intensive analysis. Workflows stitch together a variety
of data manipulation activities such as data movement, data
transformation or data visualization to serve the goals of the
scientific study. The stitching is realized by the constructs
made available by the workflow system used and is largely
shaped by the environment in which the system operates and
the function undertaken by the workflow.
A variety of workflow systems are in use [10] [3] [7] [2]
serving several scientific disciplines. A workflow is a software
[14] and CrowdLabs [8] have made publishing and finding
workflows easier, but scientists still face the challenges of re-
use, which amounts to fully understanding and exploiting the
available workflows/fragments. One difficulty in understanding
workflows is their complex nature. A workflow may contain
several scientifically-significant analysis steps, combined with
various other data preparation activities, and in different
implementation styles depending on the environment and
context in which the workflow is executed. The difficulty in
understanding causes workflow developers to revert to starting
from scratch rather than re-using existing fragments.
Through an analysis of the current practices in scientific
workflow development, we could gain insights on the creation
of understandable and more effectively re-usable workflows.
Specifically, we propose an analysis with the following objec-
tives:
1) To reverse-engineer the set of current practices in work-
flow development through an analysis of empirical evi-
dence.
2) To identify workflow abstractions that would facilitate
understandability and therefore effective re-use.
3) To detect potential information sources and heuristics
that can be used to inform the development of tools for
creating workflow abstractions.
In this paper we present the result of an empirical analysis
performed over 177 workflow descriptions from Taverna [10]
and Wings [3]. Based on this analysis, we propose a catalogue
of scientific workflow motifs. Motifs are provided through i)
a characterization of the kinds of data-oriented activities that
are carried out within workflows, which we refer to as data-
oriented motifs, and ii) a characterization of the different man-
ners in which those activity motifs are realized/implemented
within workflows, which we refer to as workflow-oriented
motifs. It is worth mentioning that, although important, motifs
Fig. 3. Distribution of Data-Oriented Motifs per domain
Fig. 3. Distribution of Data-Oriented Motifs per domain Fig. 5. Data Preparation Motifs in the Genomics Workflows

We do this repeatedly for the same datasets

Top Down: Big Micro Data(sets)
• North Atlantic Population Project (NAPP)
• Integrated Public Use Microdata Series (IPUMS)
• Mosaic

Top Down: Big Micro Data(sets)
• North Atlantic Population Project (NAPP)
• Integrated Public Use Microdata Series (IPUMS)
• Mosaic
• Only data slices can be downloaded
• Standardisation leads to loss of detail
• Results are not mutually compatible
• Large scale efforts are very expensive

Top Down: Big Micro Data(sets)
• North Atlantic Population Project (NAPP)
• Integrated Public Use Microdata Series (IPUMS)
• Mosaic
• Only data slices can be downloaded
• Standardisation leads to loss of detail
• Results are not mutually compatible
• Large scale efforts are very expensive
… and they do not solve the problem!

… the current workflow

… the current workflow

… the current workflow
Do adverse conditions (Great Depression) around birth or
early in life affect socioeconomic and health outcomes?

… the current workflow
Do adverse conditions (Great Depression) around birth or
early in life affect socioeconomic and health outcomes?
Does GDP per capita at birth year negatively
affect occupational status in later life?

… the current workflow
Do adverse conditions (Great Depression) around birth or
early in life affect socioeconomic and health outcomes?
Dutch “Hunger-winter” studies (cf Lindeboom)
Does GDP per capita at birth year negatively
affect occupational status in later life?

… the current workflow
Do adverse conditions (Great Depression) around birth or
early in life affect socioeconomic and health outcomes?
Thomasson and Fishback. 2014. “Hard Times in the Land of Plenty: The Effect on Income and
Disability Later in Life for People Born during the Great Depression.” Expl in Eco Hist 54: 64–78.
Dutch “Hunger-winter” studies (cf Lindeboom)
Does GDP per capita at birth year negatively
affect occupational status in later life?

… the current workflow
bryr AGE OCCHISCO hiscocode hiscam gdppc
1870 21 98560 9-85.55 48.70 1694.525258
1870 21 99120 9-99.10 47.88 1694.525258
1873 18 53220 5-32.10 51.65 1841.878773
1870 21 13210 1-30.00 77.29 1694.525258
1873 18 54010 5-40.90 53.27 1841.878773
1874 17 61110 6-11.10 52.61 1853.715852

… the current workflow
bryr AGE OCCHISCO hiscocode hiscam gdppc
1870 21 98560 9-85.55 48.70 1694.525258
1870 21 99120 9-99.10 47.88 1694.525258
1873 18 53220 5-32.10 51.65 1841.878773
1870 21 13210 1-30.00 77.29 1694.525258
1873 18 54010 5-40.90 53.27 1841.878773
1874 17 61110 6-11.10 52.61 1853.715852
1. Gather and enter own data
2. Find data on multiple repositories
3. Download
4. Clean and reshape
5. Merge
6. Clean and reshape…
7. Analyse

… the current workflow
bryr AGE OCCHISCO hiscocode hiscam gdppc
1870 21 98560 9-85.55 48.70 1694.525258
1870 21 99120 9-99.10 47.88 1694.525258
1873 18 53220 5-32.10 51.65 1841.878773
1870 21 13210 1-30.00 77.29 1694.525258
1873 18 54010 5-40.90 53.27 1841.878773
1874 17 61110 6-11.10 52.61 1853.715852
Link occupations in census micro data…
… to standardised occupations …
… to appropriate occupational status scores …
… to country level GDP at birth year
1. Gather and enter own data
2. Find data on multiple repositories
3. Download
4. Clean and reshape
5. Merge
6. Clean and reshape…
7. Analyse

… the current workflow

… the current workflow

… the current workflow

… the current workflow

… the current workflow
Not a very complicated research question…

… the current workflow
Not a very complicated research question…
… only one sample …

… the current workflow
Not a very complicated research question…
… only one sample …
What if we want to answer more involved questions?

"Studies that have plotted data set size against the number of data sources reliably uncover a skewed
distribution. Well-organized big science efforts featuring homogenous, well-organized data represent only a
small proportion of the total data collected by scientists. A very large proportion of scientific data falls in the
long-tail of the distribution, with numerous small independent research efforts yielding a rich variety of specialty
research data sets. The extreme right portion of the long tail includes data that are unpublished; such as siloed
databases, null findings, laboratory notes, animal care records, etc. These dark data hold a potential wealth
of knowledge but are often inaccessible to the outside world."

In the fast moving data analysis industry, real-time traceability
could help identify supply chain, brand and repetitional risks

Our Goals
• Empower individual researchers to
• Code and harmonize individual datasets according to best practices of the community
(e.g. HISCO, SDMX, World Bank, etc.) or against their colleagues
• Share their own code lists with fellow researchers
• Align code lists across datasets
• Publish their standards-compliant datasets
• Perform analyses across multiple datasets at the same time
• While tracking provenance of both data and analyses

A Linked Data Handbook for Historians? Nah…

Exists
Frequency Table
Variable does not yet existVariables
Mappings
Publish
Augment
Includes both external Linked Data and
standard vocabularies, e.g. World Bank
External (Meta) Data
Existing Variables
& Codes
Provenance tracking of all data
External Datasets
Structured Data Hub
legenddatalegenddata

Exists
Frequency Table
Variable does not yet existVariables
Mappings
Publish
Augment
Includes both external Linked Data and
standard vocabularies, e.g. World Bank
External (Meta) Data
Existing Variables
& Codes
Provenance tracking of all data
External Datasets
Structured Data Hub
legenddatalegenddata
Linked Statistical Dimensions

Dedicated Pipelines
NAPP

surname age occupation sex
Fumes 20 cigar maker female
Bridges 45 civil engineer female
Moves 17 dancer male

surname age occupation sex
Fumes 20 cigar maker female
Bridges 45 civil engineer female
Moves 17 dancer male
achternaam leeftijd beroep geslacht
Fumes 20 sigarenmaker v
Bridges 45 ingenieur v
Moves 17 danser m

surname age occupation sex
Fumes 20 cigar maker female
Bridges 45 civil engineer female
Moves 17 dancer male
achternaam leeftijd beroep geslacht
Fumes 20 sigarenmaker v
Bridges 45 ingenieur v
Moves 17 danser m

surname age occupation sex
Fumes 20 cigar maker female
Bridges 45 civil engineer female
Moves 17 dancer male
achternaam leeftijd beroep geslacht
Fumes 20 sigarenmaker v
Bridges 45 ingenieur v
Moves 17 danser m

surname age occupation sex
Fumes 20 cigar maker female
Bridges 45 civil engineer female
Moves 17 dancer male
achternaam leeftijd beroep geslacht
Fumes 20 sigarenmaker v
Bridges 45 ingenieur v
Moves 17 danser m
achternaam leeftijd beroep sdmx:Sex
Fumes 20 sigarenmaker sdmx:F
Bridges 45 ingenieur sdmx:F
Moves 17 danser sdmx:M
surname age occupation sdmx:Sex
Fumes 20 cigar maker sdmx:F
Bridges 45 civil engineer sdmx:F
Moves 17 dancer sdmx:M

surname age occupation sex
Fumes 20 cigar maker female
Bridges 45 civil engineer female
Moves 17 dancer male
achternaam leeftijd beroep geslacht
Fumes 20 sigarenmaker v
Bridges 45 ingenieur v
Moves 17 danser m
achternaam leeftijd beroep sdmx:Sex
Fumes 20 sigarenmaker sdmx:F
Bridges 45 ingenieur sdmx:F
Moves 17 danser sdmx:M
surname age occupation sdmx:Sex
Fumes 20 cigar maker sdmx:F
Bridges 45 civil engineer sdmx:F
Moves 17 dancer sdmx:M

Utrecht 1829 Utrecht 1839

Utrecht 1829 Utrecht 1839

An ecosystem is a community of living organisms in conjunction
with the nonliving components of their environment (things like
air, water and mineral soil), interacting as a system.
- Wikipedia

… the current workflow
Does GDP per capita at birth year negatively
affect occupational status in later life?
●
●
●
●
●
●
●
●
●
●
●
●●
●
20 30 40 50 60 70
3.984.004.024.04
Canada
age
log(hiscam)
●
●
●
●
●
●
●
●
●
●
●
●
●
●
6.8 7.0 7.2 7.4
3.984.004.024.04 Canada
log(gdppc)
log(hiscam)
log(hiscam) log(hiscam)
(Intercept) 4.420*** 3.616***
(0,039) (0,134)
log(gdppc) -0.058*** 0.036**
(0,005) (0,018)
I(age^2) -0.000***
0,000
age 0.007***
0,000
R2 0,003 0,013
Adj. R2 0,003 0,012
Num. obs. 36201 36201
RMSE 0,142 0,142

… the current workflow
Does GDP per capita at birth year negatively
affect occupational status in later life?
●
●
●
●
●
●
●
●
●
●
●
●●
●
20 30 40 50 60 70
3.984.004.024.04
Canada
age
log(hiscam)
●
●
●
●
●
●
●
●
●
●
●
●
●
●
6.8 7.0 7.2 7.4
3.984.004.024.04 Canada
log(gdppc)
log(hiscam)
log(hiscam) log(hiscam)
(Intercept) 4.420*** 3.616***
(0,039) (0,134)
log(gdppc) -0.058*** 0.036**
(0,005) (0,018)
I(age^2) -0.000***
0,000
age 0.007***
0,000
R2 0,003 0,013
Adj. R2 0,003 0,012
Num. obs. 36201 36201
RMSE 0,142 0,142
Identify locally, extrapolate globally?

… the new workflow
Does GDP per capita at birth year negatively
affect occupational status in later life?

… the new workflow
Does GDP per capita at birth year negatively
affect occupational status in later life?
1. Discover data on datalegend
2. Explore
3. Build or reuse a query
4. Analyse

… the new workflow
Does GDP per capita at birth year negatively
affect occupational status in later life?
1. Discover data on datalegend
2. Explore
3. Build or reuse a query
4. Analyse
http://data.socialhistory.org/resource/napp/OCCHISCO/54020

… the new workflow
Does GDP per capita at birth year negatively
affect occupational status in later life?
1. Discover data on datalegend
2. Explore
3. Build or reuse a query
4. Analyse
http://data.socialhistory.org/resource/napp/OCCHISCO/54020
http://yasgui.org

… the new workflow
Does GDP per capita at birth year negatively
affect occupational status in later life?
1. Discover data on datalegend
2. Explore
3. Build or reuse a query
4. Analyse
http://data.socialhistory.org/resource/napp/OCCHISCO/54020
http://yasgui.org
http://grlc.clariah-sdh.eculture.labs.vu.nl/clariah/wp4-queries/api-docs

… the new workflow
Does GDP per capita at birth year negatively
affect occupational status in later life?

… the new workflow
Does GDP per capita at birth year negatively
affect occupational status in later life?

… the new workflow
canada sweden
(Intercept) 3.616*** 4.430***
(0,134) (0,033)
log(gdppc) 0.036** -0.070***
(0,018) (0,004)
I(age^2) -0.000*** -0.000***
0,000 0,000
age 0.007*** 0.001***
0,000 0,000
R2 0,013 0,021
Adj. R2 0,012 0,021
Num. obs. 36201 275127
RMSE 0,142 0,102
●
●
●
●
●
●
●
●
●
●
●
●●
●
20 30 40 50 60 70
3.984.004.024.04
Canada
age
log(hiscam)
●
●
●
●
●
●
●
●
●
●
●
●
●
●
6.8 6.9 7.0 7.1 7.2 7.3 7.4 7.5
3.984.004.024.04
Canada
log(gdppc)
log(hiscam)
●
●
●
●●
●
●
●
●●●●
●●
●
●
●
●
●
●
●
●
●●
●
●●
●●●●
●
●
●
●
●●●●
●
●
●
●●●
●
●●●
●
●
●
●
●
●●
●
●
20 30 40 50 60 70
3.903.943.984.02
Sweden
age
log(hiscam)
●
●●●
●
●●●
●●
●
●●●
●
●
●
●●●●
●
●●
●
●
●
●
● ●
●
●●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
● ●
●
●
6.8 6.9 7.0 7.1 7.2 7.3
3.903.943.984.02
Sweden
log(gdppc)
log(hiscam)
Does GDP per capita at birth year negatively
affect occupational status in later life?

… the new workflow
Does GDP per capita at birth year negatively
affect occupational status in later life?

… the new workflow
Does GDP per capita at birth year negatively
affect occupational status in later life?

… the new workflow
Does GDP per capita at birth year negatively
affect occupational status in later life?

Discussion
• Data-driven research in the humanities is too expensive and confined to single
datasets.
• Linked Data can be a solution, but historians cannot be expected to change
their current workflow, or craft RDF by hand.
• QBer allows historians to upload their data, connect it to earlier work by peers,
while preserving provenance of their steps.
• The inspector view gives instant feedback of the impact on the network
• Standard SPARQL queries are converted to APIs through grlc.
• Research questions can thus be shared, replicated and applied to new data.
• This gives rise to different roles of researchers in our ecosystem
legenddatalegenddata

Discussion
• Data-driven research in the humanities is too expensive and confined to single
datasets.
• Linked Data can be a solution, but historians cannot be expected to change
their current workflow, or craft RDF by hand.
• QBer allows historians to upload their data, connect it to earlier work by peers,
while preserving provenance of their steps.
• The inspector view gives instant feedback of the impact on the network
• Standard SPARQL queries are converted to APIs through grlc.
• Research questions can thus be shared, replicated and applied to new data.
• This gives rise to different roles of researchers in our ecosystem
legenddatalegenddata