Map the GLAM

Map the GLAM is part of a PhD thesis focusing on cultural content aggregators, such as Wikimedia Commons and Europeana. In particular, Map the GLAM visualises the metadata of digital images uploaded by ETH-Bibliothek to Wikimedia Commons. The thesis aims to discover interface characteristics that may foster the access and usage of digital images released under open licenses by Galleries, Libraries, Archives and Museums (GLAMs). The project has two main goals. On the one hand, it tries to define a set of visual representations to analyse the status and the spread of a digitised collection in Wikipedia and its sister projects. On the other hand, it tries to identify usability issues that hinder the access and usage use of digital images.

The source code of the project is released under GNU General Public License and available on GitHub.

Map the GLAM is a project by Giovanni Profeta. It is developed within DensityDesign research lab in the framework of the PhD in Design at Politecnico di Milano, with the support of the Laboratory of visual culture SUPSI.

Foundations

Cultural content aggregators, such as Wikipedia, Europeana and European Library are complex information systems collecting digitised collection coming from multiple cultural institutions. They have a significant impact on culture and society.^[1]

Because of the broad audience, there is a growing interest of GLAMs in cooperating with these web platforms to make their digitised collections open and largely accessible. Cultural content aggregators are relevant because of the opportunity to find digitised items coming from multiple digital archives in one platform that can be used with few restrictions. Cultural aggregators are becoming more technologically efficient and cultural institutions are adopting rigorous sharing methods for their digital resources. Nevertheless, the user interfaces of these web platforms show some usability issues. Thus, the majority part of the available heritage is invisible to the end-user. ^[2]

This project targets both GLAMs and end-users of cultural content aggregators. On the one hand, Map the GLAM tries to identify more advanced metrics and visualization models to analyse the online impact of a digitised collection. On the other hand, a novel cultural content aggregator (called GLAM Culture Hub) was designed to evaluate possible solutions to fix current usability issues for end-users.

Map the GLAM is based on the results achieved through the GLAM visual tool research project, the analysis of the most widely used GLAM tools (such as Baglama and GLAMorous), and an in-depth analysis of the most relevant cultural content aggregators in Europe.

Research method

The project consists of three main steps: data gathering, data visualisation and design of a novel cultural content aggregator. Before the project started, a map of the GLAMs contributing to Wikimedia Commons was made to select a specific cultural institution to be investigated. Among them, ETH-Bibliothek is chosen because it is one of the leading cultural institutions in terms of files uploaded in Wikimedia Commons. Furthermore, its photographic collections have been used in several projects and linguistic versions.

Visualisation protocols

A set of visualisation protocols describe the steps to be performed from data gathering to data visualisation. It aims to make the data visualizations easy to replicate.

Map of the GLAMs

The following are the steps conducted to make the map of the GLAMs:

Gathering the list of the GLAMs, and the related number of files uploaded, starting from the following categories on Wikimedia Commons with 3 degrees of nesting:
GLAMs with less than 50 files uploaded are excluded
Gathering the coordinates of the cultural institutions from Open Street Map
Visualising the map by using D3.js and Leaflet, with tiles by Open Street Map.

Images' upload

The following are the steps conducted to gather and visualise data about the pictures uploaded by ETH Library on Wikimedia Commons.

Gathering the list of all the images uploaded by ETH Library (via Wikimedia Commons API)
Gathering the upload date and the license of the photos (via Wikimedia Commons API)
Visualising the timeline by using D3.js.

Images' size

The following are the steps conducted to gather and visualise data about the size of images uploaded by ETH Library on Wikimedia Commons.

Starting from the list of all the images uploaded by ETH Library (see Images' upload), gather the digital size of the pictures (via Wikimedia Commons API)
Gathering the physical image size by scraping the Wikimedia Commons page
Parsing of the data about the physical image size to make them easy to compare with the digital size
Visualising the digital size and the physical size, through a heatmap, by using D3.js.

Images' usage

The following are the steps conducted to gather and visualise data about images usage over Wikipedia and Wikimedia Commons.

Starting from the list of all the images uploaded by ETH Library (see Images' upload), gather the list of pages including the images by scraping the Wikimedia Commons page
Collecting all the revisions of the pages including the ETH Library images via Wikimedia Commons API) and Wikipedia API
Parsing all the reviews to check the date in which the ETH Library image was included in the page
Visualising the Wikipedia and Wikimedia Commons timelines by using D3.js.

Features of used images

The following are the steps conducted to gather and visualise data about the images used in Wikipedia and Wikimedia Commons.

Resuming all the revisions of the pages including the ETH Library pictures (see Images' usage) and the metadata of the images (see Images' size)
Parsing of the data to obtain the list of images used in Wikipedia and Wikimedia Commons
Parsing of the data to collect sets of authors, dates, mediums and orientation.

Pageviews

The following are the steps conducted to gather and visualise the daily pageviews of pages containing ETH Library images.

Starting from the list of pages including ETH Library images (see Images' usage), gather the daily page views via Wikimedia API
Parse the data gained to obtain the average daily page views
Visualising the chart by using D3.js.

Images' position within pages

The following are the steps conducted to gather and visualise the data about the Images' position within Wikipedia pages.

Starting from the list of Wikipedia pages including ETH Library images (see Images' usage), gather the HTML file of the page.
Parsing of the HTML file to obtain the starting tag of the ETH Library image as a relative value, from 0 (the very beginning of the HTML page) to 100 (at the end of the page). In this step, data about the page typology (article/user page/discussion page/other types of pages) are also collected.
Gathering of minimum, maximum and median values for every page typology.

This method does not collect data about the visual position of an image within a page, but, after some manual checks, I found that it provides accurate results.

Visualisations

Map of the GLAMs

The maps show the GLAMs (galleries, libraries, archives, museums, and other cultural institutions) contributing to Wikimedia Commons. The map includes 168 GLAMs: 3 are galleries, 32 are libraries, 26 are archives, 50 are museums, and 57 are other cultural institutions (foundations, public agencies etc.)

List of GLAMs contributing to Wikimedia Commons.
Galleries contributing to Wikimedia Commons.
Libraries contributing to Wikimedia Commons.
Archives contributing to Wikimedia Commons.
Museums contributing to Wikimedia Commons.
Other cultural institutions contributing to Wikimedia Commons.

Visualising the contribution of ETH Library

Images' upload

The timeline (from May 2016 to June 2018) shows the uploads of ETH Library images in Wikimedia Commons. The data were scraped through the Commons API by using the category “Media contributed by the ETH-Bibliothek”.

Images' size

The visualisation shows the digital and the physical size of ETH Libraries images. From the comparison, it looks like there is sometimes a swap of width and height in the physical size of the images (this might be due to the unstructured field "Dimension" in the images'metadata).

Images' usage

The visualisation shows the typologies of pages and linguistic version in which collection pictures were added.

Features of used images

The visualisation shows the metadata of the photos that have been added to Wikipedia and Commons pages. Wikipedia pie charts refer to 5% of the collection. Commons pie chart refers to 50% of the collection.

Pageviews

The visualisation shows the daily page views of Wikipedia and Commons pages containing ETH Library images. The average is calculated between July 2016 and July 2018.

Images' position within pages

The visualisation shows the average position of ETH Library images within Wikipedia pages. The data are gathered through the analysis of the wikitext. It is not based on the visible position within the page).

Timeline of images upload.
Digital and the physical size of ETH Libraries images.
Timeline of the usage of ETH Libraries images in Wikipedia pages.
ETH Libraries images in Wikimedia Commons pages.
Charts of the feature of the images used in Wikipedia and in Wikimedia Commons.
Chart of the daily page views of pages containing ETH Library images.
Chart of the position of the ETH Libraries images within Wikipedia pages.

Most used images

The following is the top 100 of the most used ETH Library images over Wikipedia and its sister projects. This list was used to search on the web, via TinEye, the websites using ETH Library images. Some of them were contacted in order to gather some information on the access and usage of the images.

Findings

The data visualisations have made visible the vast extension of the ETH collection regarding time, authors and mediums and a relatively low use within Wikimedia projects. Over the years 2016-2017-2018, 8.800 users contribute to add around 1.000 images in more than 1.000 Wikipedia pages. The photos were initially added to the German version of Wikipedia articles. Then in French, English and other 43 linguistic versions. The most used images portray intellectuals and other famous people but also aerial views of cities. The images are mainly added to pages about European and African cities. During summer 2017, there was a pick in the usage of pictures because there was a Wikipedian in residence at ETH library. After an increase in the usage of 400% in one month, the usage keeps growing at a rate of 20% in the following months. As expected, the page views follow a long tail pattern. In Wikimedia Commons, the head is much thicker and lower than in Wikipedia (very few pages are popular). The page views on Commons are much less relevant than those on Wikipedia (about 10% of those on Wikipedia, although there are 94% more pages containing files on commons). It was discovered that during the summertime (June-September) there is an evident decrease in the page views.

GLAM Culture Hub

GLAM Culture Hub is an interactive mockup of a digital archive collecting digitised items coming from multiple GLAMs (also called cultural content aggregator). It was designed after the analysis of the spread of ETH Library's images over Wikipedia and its sister projects and the analysis of existing cultural content aggregators. The goal of GLAM Culture Hub is to evaluate, through an online survey, interface characteristics that may foster the access and usage of digital images within content aggregators.

User interface

The preliminary research, conducted through the benchmark and user survey, showed that current cultural content aggregators have several usability issues regarding content, item classification and user interface. Among the content issues, there is a lack of information and unclear terms of use. The classification is sometimes misleading. And often the user interfaces lack tools to explore the digitised collections a consistent design. GLAM Culture Hub try to propose some design guidelines to fix these usability issues.

It consists of three main pages: the homepage, the list of items and the single item. Three types of access to the platform were conceived: unregistered users, registered users, registered users working for a GLAM.

GLAM Culture Hub features content coming from Wikimedia Commons and Europeana.

Interactive prototype and survey

After watching the video demo or navigating the interactive prototype, you can fill in the survey. The survey will take 10 minutes to complete:

> Evaluation of the GLAM Culture Hub interface

At the following link, you can find the interactive prototype. N.B.: the interactive prototype consists of static images with the addition of some sensible areas; thus, functions cannot be performed.

> GLAM Culture Hub Interactive prototype