krishanr / arxiv-explore

This project uses the arxiv metadata-v5 to create an interactive dashboard for viewing influential preprints.

This project uses the arxiv metadata-v5 to create an interactive dashboard for viewing influential preprints.

Once installed, the dash app, can be run locally using the commands:

$ conda activate arxivexplorer
$ python app.py

Also see the jupter notebook for scratch work done before the dash app was made.

Installation

First create the conda environment using:

$ conda env create --file=environment.yml
$ conda activate arxivexplorer

Then once the arxivexplorer enviroment is activated create the web data using:

$ make build_features

Finally, run the dash app locally using:

$ python app.py

The app can also be run on heroku by following the instructions.

Data

This app uses the arxiv metadata-v5 available from the Google cloud, at "gs://arxiv-dataset/metadata-v5". Specifically, the files "arxiv-metadata-oai.json" and "internal-citations.json" are used.

References

• https://www.kaggle.com/steubk/arxiv-taxonomy-e-top-influential-papers?rvi=1