'NA' strings in the source MITMA CSV files caused DuckDB conversion errors when casting to enums or other non-string types. This fix handles 'NA' values in columns like renta, distancia, residencia, origen, destino, actividad, and others across v1 and v2 data.spod_get_zonestime_slot column has been removed from the output of spod_get() and spod_convert(). Please use the hour column instead. Deprecation message removed.spanishoddata paper is out: Kotov, E., Vidal-Tortosa, E., Cantú-Ros, O. G., Burrieza-Galán, J., Herranz, R., Gullón Muñoz-Repiso, T., & Lovelace, R. (2026). spanishoddata: A package for accessing and working with Spanish Open Mobility Big Data. Environment and Planning B: Urban Analytics and City Science. DOI:10.1177/23998083251415040{target='_blank'}
Package now has comprehensive unit tests coverage.
Tests on live data run weekly on GitHub workers to alert maintainers of any issues.
spanishoddata package and paper are now featured in DuckDB Library
spod_set_data_dir() function to set the data directory instead of advising to set the environment variable manually.spod_quick_get_zones() is a new function to quickly get municipality geometries to match with the data retrieved with spod_quick_get_od() (PR #163). This function is experimental, just as the spod_quick_get_od() function, as the API of the Spanish Ministry of Transport may change in the future. It is only intended for quick analysis in educational or other demonstration purposes, as it downloads very little data compared to the regular spod_get_od(), spod_download() and spod_convert() functions. The requests are cached in memory of the current R session with memoise package, so repeated calls to spod_quick_get_zones() will not cause repeated requests to the API and will allow the user to get the data faster from repeat calls.
Experimental spod_check_files() function allows to check consistency of downloaded files with Amazon S3 checksums (PR #165). ETags for v1 data are stored with the package, and for v2 data they are fetched from Amazon S3. The checks may fail for May 2022 data and for some 2025 data, as the remote cheksums that are used for checking the file consistency are incorrect. We are working on solving this in future updates, for now, kindly rely on the built-in file size checks of spod_download(), spod_get(), and spod_convert().
spod_get() and spod_convert() are now up to x100,000 faster when you have all (or a lot of) data downloaded, but only requesting several days in the call to spod_get() or spod_convert(). This is thanks to a new smarter filtering strategy (issue #159, PR #166).
Metadata is now fetched from Amazon S3 storage of the original data files, which allows validation of downloaded files (issue #126) with both size and checksum. PR #165.
Metadata fetched by spod_available_data() has extra columns such as data type, zones and period, see help ?spod_available_data() for details.
Memory allocation is now delegated to DuckDB engine, which by default uses 80% of available RAM. Beware that in some HPC environments this may detect more memory than is actually available to your job, so set the limit manually to 80% of RAM available to your job with max_mem_gb argument of spod_get(), spod_convert(), spod_connect() functions. This will also improve performance in some cases, as DuckDB is more efficient than R at memory allocation (PR #167).
More reliable, but still multi-threaded data file downloads using base R utils::download.file() instead of curl::multi_download() which failed on some connections (issue #127), so now curl dependency is no longer required. PR #165.
spod_quick_get_od() is working again. We fixed it to work with the updated API of the Spanish Ministry of Transport (PR #163, issue #162). This function will remain experimental, just as the spod_quick_get_zones() function, as the API of the Spanish Ministry of Transport may change in the future. It is only intended for quick analysis in educational or other demonstration purposes, as it downloads very little data compared to the regular spod_get_od(), spod_download() and spod_convert() functions. The requests are cached in memory of the current R session with memoise package, so repeated calls to spod_quick_get_od() will not cause repeated requests to the API and will allow the user to get the data faster from repeat calls.
spod_convert() now accepts overwrite = 'update' with save_format = 'parquet' (#161) previously it failed because of the incorrect check that asserted only TRUE or FALSE (#160)
spod_cite() function to easily cite the package and the data (#134)hour column is superseeded by time_slot column in the output of spod_get() and spod_convert(). time_slot is deprecated. It is still present in the tables, but will be removed in the end of 2025 but going forward please use the new hour column. Otherwise it is exactly the same as before, this is just a name change. (#132)spod_quick_get() does not rely on metadata download anymore and can be used without setting the data directory with spod_set_data_dir() (and therefore does not cause a warning if the data directory is not set).
hour (ex-time_slot) column is now right next to the date column in the output of spod_get() and spod_convert() (#)
maximum available CPU cores check is now turned off to improve compatibility when running the package from within a container in high performance computing environments (see #130 and #140 for details)
minor documentation improvements and updates
minor bug fixes