Meeting 2016-05-04

[twhyntie]

Notes from OpenCosmics (ad-hoc) meeting

• Time: Wednesday 4th May 2016, 1pm
• Location: CERN R1
• Present: @RaoOfPhysics @ozel @pingud98 @twhyntie

Reminder of the project goals and scope

• Define a common data format for open cosmic ray projects to enable project interoperability;
• Create a repository (possibly with the help of CERN) for the data stored in this format (and possibly raw data where appropriate/necessary);
• Mozilla Science Lab involvement through Collaborate.

Potential data sources

(NB: listing a project here does not imply endorsement or official involvement!)

• Crayfis;
• Cosmic Pi;
• HiSpark;
• ERGO;
• QuarkNet;
• CERN@school;
• Nebraska desert-based detector network.

The Open Cosmic data format

• It was noted that there will probably need to be different levels of data available, e.g. RAW, RECO (reconstructed particle info), etc.
• Some projects only make data available at the event level anyway;
• RAW output from detectors could be uploaded to a separate public repository (with DOI) anyway and integrated later. This would enable immediate progress and provide test cases for the data specification draft.
• Not just detector data required - environmental data should be recorded/obtained too, though this could be from some other source (which could be verified/validated by local data too);
• Would be useful to have a physics-based use case to test formats (and conversion of RAW formats) on. Could @tpmccauley help with this as a cosmic ray expert?

Update on Cosmic Pi status and specs

• @pingud98 provided an update on the Cosmic Pi status;
• Testing prototype 1.2
• 30 units to be delivered soon in time to be prizes for the runners-up of the CERN Beamline competition;
• 100 nanosecond timing resolution;
• Environmental data: temperature, pressure, GPS, magnetic field, detector orientation - all from one box!
• Data provided as a bitstream from an Arduino, passed Python-wrapped to a Raspberry;
• Data is event-based, assuming no more than 5/s. Each event tagged with a timestamp and environmental data;
• Data format based on Python dictionary/JSON - using Mongo DB for data storage;
• Planned project duration: 5 years starting September 2016;
• Licenses: CERN Open Hardware, GPL v2 for software, CC0 for data;
• No personalised data taken or stored to avoid privacy issues.

Discussion of Safecast

• A project that uses a GM tube + data recorder box to map radiation (counts per second), setup post-Fukushima;
• @ozel showed us his detector and the maps, plus the API;
• Data recorded (i.e. count/s) probably not useful for cosmics measurement but great example of a successful project.

CERN@school

• @twhyntie mentioned CERN@school - not a dedicated cosmic ray project (very small active sensor element) but the Timepix detector is capable of recording cosmic ray particles in the course of other measurements;
• Makes sense to at least ensure Timepix data compatible with other projects to facilitate new avenues of research (particularly with the Timepix detectors on the LUCID satellite and the International Space Station);
• @pingud98 raised interesting point of GPS integration with the Timepix measurements - not currently done with dedicated equipment. Something for CERN@school to think about (especially if Cosmic Pi boxes available...).

Organisation

• @RaoOfPhysics suggested using GitHub to organise the project - excellent idea!

Action items

• We should draft a physics-based use case upon which to help define what the data format would need to include;
• We should draft a standard for the format based on RSS, FITS, etc.
• Anyone involved with projects that have sample data available should upload them to a repository to facilitate testing of the format and any conversion software etc.

[twhyntie]

If I've missed anything, feel free to edit/add stuff as appropriate. Thanks again all!

[RaoOfPhysics]

Paging @tpmccauley.

[Tontonis]

Query about action 3: If data from a project is available but noone from it currently involved (like HiSPARC) are we cautious about using their data or shall we just go for it? I've been using some of HiSPARC's data for looking at the CosmicPi analysis and for evaluating data formats but only on my own machine so far. Am happy to stick it up somewhere and deal with any bad feedback.

[twhyntie]

@Tontonis Do we know the license the data is issued under? The CERN@school sample set is CC-BY, but I believe the ultimate aim for Open Cosmics is CC0.

It would be very useful to have the set uploaded to a citable repository like Zenodo or FigShare. Whether this should/can/must be done by a HiSPARC team member would (I guess?) probably depend on the license and any collaboration agreements signed... what do you think?

[RaoOfPhysics]

Paging @153957. :D (Been meaning to get in touch with you for a while, Arne. Sorry it's happened so late.)

See https://github.com/HiSPARC/publicdb and http://data.hisparc.nl/show/stations_by_country/

[153957]

Our (HiSPARC) data is licensed under Creative Commons Attribution-ShareAlike 4.0.

We have made some efforts recently to have station maintainers measure the actual locations of their detectors relative to the GPS antenna (whos position we automatically get). This allows us to do direction reconstruction with some individual detection stations if they have four detectors.
Currently we are working on implementing automated reconstructions for coincidences between multiple stations, for which we are now calibrating GPS timing offsets. The reconstructions will be part of the downloadable data. - This means that our data files are likely to contain more data in the future.