Planet Hunters Analysis Database


About Planet Hunters Analysis Database

The Planet Hunters Analysis Database (PHAD) is an experimental new service at MAST developed in partnership with the Planet Hunters Zooniverse team. PHAD receives data in real time from the Planet Hunters project as members of the Zooniverse community analyze the data from TESS and as potential new exoplanet transits are discovered.

Fig. 1 - A screengrab of the Planet Hunters interface, together with a number of potential transits marked. Transit markings from a collection of Zooniverse volunteers are combined to produce the results you see here on the PHAD site.

What are the values in the table?

There are two views on the PHAD site for essentially the same data 1) a ‘transit’ oriented view listing the subject_ids (the data identifier from Planet Hunters), tic_id, and the centers of the transit box and 2) a ‘user’ oriented view which lists every subject-transit-user combination where the analyses from multiple individuals has resulted in a potential transit identification.

The tables on this website don’t list every single classification from the Planet Hunters website (although they are available on the bulk downloads pages), instead, what we’re displaying here are transit candidates when multiple volunteers from the Zooniverse project have agreed sufficiently that the clustering algorithm implemented by the Zooniverse team has identified a transit candidate.

How can I search for my classifications?

You can search for your potential transit candidates by clicking on the ‘User’ tab, and entering either your Zooniverse login name or credited name.

How are the candidates in this table generated?

The algorithm used here is the work of the Planet Hunters TESS project lead, Nora Eisner, a PhD candidate at the University of Oxford. The table includes all subjects were at least one transit has reached a weighted_count greater than 5; all possible transits for that subject are reported.

The logic behind the transit identification is as follows:

  • The DBSCAN algorithm analyzes all of the transit markings (boxes drawn in the Planet Hunters user interface) for a single subject (light curve) and combines them to form a ‘cluster’
  • The average position (‘Center’ and ‘Width’ columns here) of all the markings in that cluster is then taken to be the location of the transit-like event..
  • The significance of the transit (‘Weighted Counts’ column here) is the sum of all the user weights of the people who marked the transit.
  • The user weightings are calculated based on whether the volunteers correctly identify the simulated data or not; if a volunteer correctly identifies a transit the user weighting increases, and if they miss one it decreases. The amount it increases or decreases by depends on how ‘hard’ or ‘easy’ the simulation is, which is calculated based on how many other volunteers correctly identified it. You can read more about this algorithm in Schwamb et al.

Fig. 2 - A screengrab of a search result.

Are the raw classifications from Planet Hunters available?

Yes. The Zooniverse team will be providing monthly database exports for all of the raw classification data.

I’d like to follow up on some of these candidates listed here, what should I do?

Most importantly, we ask that you credit the Zooniverse volunteers listed in the table here (please use the ‘Credited Name’ entries for a given subject id or TIC id). In addition, please credit the Planet Hunters team.

For more context and discussion about a given subject, clicking the subject id field values will take you to the Planet Hunters Talk page where you can find out more information about the star being observed. If you’d like to retrieve the data at MAST for a particular TIC id, you can search the MAST portal, for example here's a search for TIC id 308454343. Finally, the team at MAST have developed a set of example Jupyter Notebooks that can be used to programmatically access the data at MAST.