Sciscinet-v2 is a refreshed update to SciSciNet which is a large-scale, integrated dataset designed to support research in the science of science domain. It combines scientific publications with their network of relationships to funding sources, patents, citations, and institutional affiliations, creating a rich ecosystem for analyzing scientific productivity, impact, and innovation. Know more.
The newer version Sciscinet-v2 is rebuilt using the latest snapshot from OpenAlex. Here are some informative comparisons of Sciscinet-v2 with Sciscinet-v1.
- How does Sciscinet-v2 compare to the previous version ?
- Does Sciscinet-v2 include precomputed metrics and linkages ?
- Do you have precomputed embeddings for papers ?
$ gsutil -m cp -r gs://sciscinet-neo/v2/embeddings/* ./path/to/your/directory/4. I donot see sciscinet_paperdetails.parquet or sciscinet_papertitleabstract.parquet on Huggingface, where can i find them ?
sciscinet_paperdetails.parquet (117 GB) and sciscinet_papertitleabstract.parquet (92GB) are hosted on Google cloud storage and Big Query
exclusively because of the size of the files. You can simply access them using the command
$ gsutil ls gs://sciscinet-neo/v2 | grep -e "sciscinet_paperdetails"
gs://sciscinet-neo/v2/sciscinet_paperdetails.parquet
$ gsutil ls gs://sciscinet-neo/v2 | grep -e "sciscinet_papertitleabstract"
gs://sciscinet-neo/v2/sciscinet_papertitleabstract.parquet5. I donot see sciscinet_journals similar to v1, where can i find journal information ?
Since Sciscinet-v2 is built on top of OpenAlex, journal information is stored in the form of Sources.
You can find the same dynamic in the file sciscinet_sources.parquet and paper-source mapping in sciscinet_papersources.parquet.
More information can be found on https://northwestern-cssi.github.io/sciscinet/.
For more information about the previous [Sciscinet-v1] version visit README_OLD.md.
