aimalz/castorpz

Experiments with CASTOR photo-z forecasting

castorpz's Issues

Data exploration: photometry

In preparation for evaluating TheLastMetric, we should check for systematic weirdness (e.g. discreteness) in the redshift-photometry space.

Data exploration: photo-z estimates

This issue can be closed when there is a Jupyter notebook that isolates the estimated photo-zs from the mock catalog and calculates the canonical statistics (bias, intrinsic scatter, cutoff in scatter by which each galaxy would be considered an outlier) on a per-galaxy basis. EDIT: We can also use the magnitude error estimates from the catalog as another dimension for selection criteria.

Once this is done we can make selection cuts on these statistics to focus on galaxies whose photo-zs are the worst without CASTOR data and then quantify the improvement in their photo-zs with CASTOR data..

Cosmology forecasting prerequisites

We need more information about the simulations to get a Fisher forecast on cosmology from n(z)s derived from samples with and without CASTOR data, specifically a covariance matrix (per tomographic bin). Then we can run it through CHIPPR's CosmoLike procedure (or a modern analogue -- strongly preferable because CHIPPR is still in Python 2, sorry).