Create a Jupyter notebook to investigate the consistency between the INSPIRE record numbers obtained from either INSPIRE or HEPData. Discrepancies usually occur because an INSPIRE record has changed its record number.

Feedback from Sabine Kraml (@sabinekraml):

The general coverage of published analyses (how many of them have HEPData entries) and how long it takes from the paper publication to the HEPData upload for the different experiments/physics groups would of course also be interesting to know, provided such information is readily at hand.

Coverage by experiment is handled by count_inspire_records_with_hepdata.ipynb.

The time from arXiv publication to HEPData release could be calculated by the difference between the created timestamp from the INSPIRE record (obtained using their REST API) and the last_updated timestamp from the first version of the HEPData record (passing URL options of format=json & light=true & version=1). A list of INSPIRE IDs could be obtained by using their REST API to search for INSPIRE records that have a corresponding HEPData record, possibly filtering by year and collaboration as in count_inspire_records_with_hepdata.ipynb. A histogram could then be plotted showing the number of papers for different time intervals. A Jupyter notebook should be produced to perform the calculations and make some plots, allowing it to easily run and adapted by other interested users.