Astronomy datasets

This module contains the implementation of multiple astronomy datasets following the tensorflow dataset API.

Currently implemented time series datasets are:

• spcc (Supernova Photometric Classification Challenge)
• plasticc (Photometric LSST Astronomical Time-series Classification Challenge)

Currently implemented image datasets are:

• slc (Strong Lens Finding Challenge)
• mirabest
• cmd (Camels MultiField Dataset)
• mlsst (Mock LSST)

This is based on the medical datasets repository https://github.com/ExpectationMax/medical_ts_datasets

Example usage

In order to get a tensorflow dataset representation of one of the datasets simply import tensorflow_datasets and this module. The datasets can then be accessed like any other tensorflow dataset.

import tensorflow_datasets as tfds
import astro_datasets

dataset, info = tfds.load(name='spcc', split='train', with_info=True)

Times series instance structure

Each instance in the dataset is represented as a nested directory of the following structure:

• static: Static variables such as photometric redshift
• static_errors: Static variable errors
• time: Scalar time variable containing the observation time
• values: Observation values of time series, these by default contain NaN for modalities which were not observed for the given timepoint.
• value_errors: Observation values of time series errors, these by default contain NaN for modalities which were not observed for the given timepoint.
• targets: Directory of potential target values, the available endpoints are dataset specific.
• metadata: Directory of metadata on an individual object

Credits

Lauren Gaughan, Samuel Gibbon