On this page

The SF Nexus is a collaborative effort by research and public libraries with science fiction collections to make SF available online, including as data. The project is based at Temple University’s Charles Library, and we want to keep growing a research community around SF collections. The current site is a prototype of what a large-scale collaborative digitization could look like across the dozens of science fiction collections in the UK and North America, including but not limited to the members of the Science Fiction Collecting Libraries Consortium.

This phase of the project shows how a library can digitize its copyrighted cultural collections and make them available as data. So far we have focused on sharing extracted features and on documenting how the corpus was deposited in HathiTrust and curated for the HathiTrust Research Center. Related work at Temple Libraries is developing local data capsules for secure, controlled computing on copyrighted corpora, along with new ways of digitizing corpora under controlled conditions.

The Paskow Science Fiction Collection

The Loretta C. Duckworth Scholars Studio has partnered with Temple University Libraries’ Special Collections Research Center (SCRC) and Digital Library Initiatives (DLI) to build a digitized corpus of copyrighted science fiction. The SCRC is known for its extensive Urban Archives, but it also holds a significant science fiction collection. The Paskow Science Fiction Collection was established in 1972, when Temple acquired 5,000 science fiction paperbacks from a Temple alumnus, the late David C. Paskow.

Later donations expanded the collection in size and in range of genres, including troves of fanzines and the papers of writers such as John Varley and Stanley G. Weinbaum. SCRC staff and undergraduate student workers compared gift titles against catalogued books and removed exact duplicates of existing holdings. A refocused collection development policy for science fiction also de-emphasized fantasy and horror, so some titles in those genres were removed as well. Over the past several years, a subset of these deduplicated science fiction titles has been digitized. That subset is the SF Nexus corpus.

Related archival collections at the SCRC can be explored in Temple’s ArchivesSpace:

The digitization project

Digitization began in 2017. The original plan is described in the Scholars Studio blog post “Building a New Wave Science Fiction Corpus”. Our Omeka S exhibit tells more of the project’s history and shows selected book covers from the collection with their metadata. Temple Now covered the work in “Byte-size books: Digitizing Temple University Libraries’ science fiction collection” by Edirin Oputu (June 15, 2020). Presentations about the project are archived on the Libraries’ Past Programs site. One example is “Science Fiction for the Future: Building Collections for Teaching and Research” with Margery Sly, Alex Wermer-Colan, and André Carrington.

Temple has deposited 348 of the scanned volumes in the HathiTrust Digital Library, where they are searchable in full text. The packaging and ingest workflow is documented on the OCR page.

Project timeline

  1. 1972

    The Paskow Science Fiction Collection is founded

    Temple acquires 5,000 science fiction paperbacks from alumnus David C. Paskow. Later gifts of fanzines and author papers (John Varley, Stanley G. Weinbaum) expand the collection over the following decades.

  2. 2017

    Digitization begins

    The Scholars Studio, the Special Collections Research Center, and Digital Library Initiatives start scanning deaccessioned duplicates from the Paskow Collection, working with a team of graduate research assistants. The plan is laid out in “Building a New Wave Science Fiction Corpus”.

  3. 2019

    Extracting features in R

    Jeff Antsen publishes an R script for extracting and structuring textual features from copyrighted corpora, the first version of the method behind the site’s datasets.

  4. 2020

    The SF Nexus at Stanford

    Alex Wermer-Colan presents “The SF Nexus: Digitizing, Curating, and Modeling Mass-Market Speculative Fiction” at Stanford’s CESTA seminar.

  5. “Byte-size books”

    Temple Now covers the project in “Byte-size books: Digitizing Temple University Libraries’ science fiction collection”.

  6. 2022

    The New Wave of Digital Collections

    Wermer-Colan and James Kopaczewski set out the project’s approach to curating copyrighted literature as data in the Transactions of the American Philosophical Society (vol. 110, no. 3). The team also presents “Digitizing and Exhibiting Copyrighted Science Fiction” at HathiTrust Community Week.

  7. The Stories We Tell Ourselves

    With Temple Arts and Humanities funding, Megan Kane and SaraGrace Stefan lead students in curating the corpus with a focus on climate fiction, producing the notebooks, datasets, and first version of this website.

  8. 2023

    Datasets and notebooks released

    Five extracted-features datasets go live on Hugging Face, alongside Python and R notebooks for sectioning, disaggregation, BookNLP, topic modeling, and Word2Vec.

  9. Toxic Water at DH2023

    Dez Miller, Alex Wermer-Colan, Megan Kane, and SaraGrace Stefan present “Modeling Eco-Poetics and Eco-Politics in 20th Century Anglophone Climate Fiction: Toxic Water” at the ADHO conference in Graz (abstract).

  10. 2025

    Data Speculations convenes

    An IMLS National Leadership Grant (LG-254864-OLS-23) to Temple and Texas A&M brings librarians, scholars, and copyright experts together for Data Speculations, a national forum on library stewardship of copyrighted contemporary culture. Ted Underwood gives the opening keynote, “Copyrighted Cultural Analytics.”

  11. 2026

    sfnexus.org

    The project moves to its own domain and reorganizes the site around the data lifecycle, from source to publication.

Our approach: collections as data

The SF Nexus is part of the wider collections as data movement. This movement asks libraries, archives, and museums to make their holdings available for computational research, not only for reading and viewing. Copyrighted twentieth-century literature is one of the hardest cases. The books that shaped modern genre fiction are often still in copyright, so the full texts cannot be shared. U.S. courts have held that building a searchable, analyzable index of copyrighted books can be fair use (Authors Guild v. HathiTrust, 2014; Authors Guild v. Google, 2015). That legal footing makes non-consumptive research possible: computational analysis that doesn’t let anyone read the works.

For an overview of our approach to curating literary texts as data, see Alex Wermer-Colan and James Kopaczewski, “The New Wave of Digital Collections: Speculating on the Future of Library Curation,” Transactions of the American Philosophical Society 110, no. 3 (2022): 211–241. Because of copyright restrictions, the SF Nexus explores speculative approaches to data curation. The goal is to share enough of each book (its extracted features) to support analysis at scale. The Scholarship page gathers the frameworks, legal guidance, and related projects this work builds on.

In the future we hope to build a combined dataset that draws on open sources of public-domain SF, such as Project Gutenberg, the Internet Archive, and Luminist.org. We also hope to add in-copyright paperbacks and pulp magazines digitized at libraries like Temple’s, shared as derived datasets like those available from HathiTrust.

Funding and support

This work was supported by Temple University Arts and Humanities grant funding for “The Stories We Tell Ourselves: The Cultural Analytics of Climate Fiction in the Age of the Anthropocene.” That project supported graduate students doing digital humanities research through the science fiction digitization project. Temple University Libraries has also invested money, staff, and resources in digitizing deaccessioned books from the Paskow Science Fiction Collection continuously since 2017.

“The Stories We Tell Ourselves” brought together undergraduate and graduate students, led by Megan Kane and SaraGrace Stefan, to curate our SF dataset with a focus on climate fiction. That work produced this website and its related materials. The curation work also led to a collaboration with Dez Miller, PhD in Comparative Literature at Emory University. This multi-institutional team compared climate fiction in Temple’s SF corpus with the SF available in HathiTrust. It then mined HathiTrust’s data capsule to study climate fiction with topic modeling and word embeddings. The team presented the results at the 2023 conference of the Alliance of Digital Humanities Organizations (DH2023, Graz) as “Modeling Eco-Poetics and Eco-Politics in 20th Century Anglophone Climate Fiction: Toxic Water.” The slides are on Zenodo.

The project has since grown into Data Speculations: A National Forum on Library Digital Stewardship for Copyrighted Contemporary Culture. The forum is funded by an Institute of Museum and Library Services National Leadership Grant for Libraries (LG-254864-OLS-23) to Temple University in collaboration with Texas A&M University. Alex Wermer-Colan is principal investigator and Sarah Potvin is co-PI. The forum’s legal team is Brandon Butler, Peter Jaszi, and Rachael Samberg. It convened in January 2025 to develop legal best practices for curating copyrighted literature. Learn more on the Data Speculations website.