Steps 5–6 · Student Research

Open research tracks

These are entry points into questions the project has not yet answered, not finished assignments. Each track names a real starting dataset, the background you need, a sensible first output and the main uncertainty you will run into. Talk to the project team before you start, so that work is coordinated and access to the research archive is arranged.

If you are new, start with Track A. Reproducing an existing result from native data is the best way to learn the chain, and it produces the project's first independently reproduced (L8) product.

Track A · Introductory–intermediate

Independent spectrum reduction

Can the 26 Aug reduced spectra, and the acceptance of 661 of 776 spectra, be reproduced from the native data with documented code?

Start with
26 Aug 2026 native spectra and its pointing manifest
Background
Python and NumPy; power spectra; fitting a baseline
Output
A notebook or script, a comparison with the archived reduced spectra, and a list of any undocumented processing steps you had to infer
Main issue
The code that made the archived products is not available, and part of the bandpass treatment depended on a stored template that is not in the archive
Track B · Intermediate

Instrument calibration

How do the C1–C4 scales compare, and how robust are G and p to the fitting method and its uncertainties?

Start with
The eight C4 reference points on the Calibration page, the 2 Sep Y-factor record and the 2 Sep antenna-temperature cube
Background
Radiometry (system temperature, Y-factor); least-squares fitting with errors in both variables
Output
A calibration note; optionally a new, separately labelled C4 re-derivation of the 2 Sep cube
Main issue
The Y-factor dropped from 4.40 dB (2 Sep) to 1.29 dB (20 Sep) for a similar warm wall, and the reason is not yet documented
Track C · Intermediate

Pointing and repeatability

How accurately and repeatably does DISH222 point, judged from the data themselves?

Start with
Repeated spectra at the same positions in the 26 Aug session; the Galactic-plane ridge in the 26 Aug and 2 Sep maps
Background
Celestial coordinates and Astropy; beam patterns
Output
Ridge-centroid and repeat-scatter measurements with an error budget, including the effect of coordinate precession
Main issue
The manifest records commanded positions, not measured ones
Track D · Intermediate–advanced

Galactic H I velocity analysis

Does a first-moment or multi-component fit recover the Galactic rotation amplitude that the peak metric compressed?

Start with
26 Aug reduced spectra (or your own reduction from Track A)
Background
Galactic rotation and kinematic distances; LSR correction; Gaussian fitting
Output
A longitude–velocity diagram and a comparison of velocity metrics against LAB
Main issue
With a 3° beam, local gas near 0 km/s blends with more distant spiral-arm gas
Track E · Intermediate

FITS and spectral-cube imaging

Can we produce standards-compliant cubes with complete coordinate and provenance metadata?

Start with
26 Aug spectral cube as a reference; your own gridding of the reduced spectra
Background
FITS and WCS; gridding and smoothing; moment maps
Output
A cube, moment maps and a proposed header standard (units, frame, epoch, calibration ID)
Main issue
Several archived FITS headers lack observation date, telescope, rest frequency, velocity frame or equinox
Track F · Advanced

Receiver and system characterization

How linear and stable is the receiver chain, and how stable is its bandpass from session to session?

Start with
The two spectrum types recorded in the 26 Aug session; new bench measurements made with the project team
Background
RF systems; software-defined radio; noise temperature
Output
A bench-test report, including an independent check of the receiver's sample format and linearity
Main issue
Needs supervised hardware access. This track does not involve operating or controlling the telescope.
Track G · Intermediate

Comparison with the LAB survey

Where do DISH222 profiles agree with LAB smoothed to a 3° beam, and where do stray radiation or the baseline pedestal dominate?

Start with
26 Aug reduced spectra; 2 Sep Galactic-centre maps
Background
Radio survey methods; beam convolution
Output
Profile-by-profile comparison and residual maps with an explanation of each systematic effect
Main issue
C4 is itself anchored to LAB, so agreement on C4 is a consistency check rather than an independent test
Track H · Long-term

Full-Galaxy reconstruction

Can a reproducible all-sky H I map be rebuilt from native sessions, on one calibration, with a documented selection?

Start with
26 Aug native spectra now; further sessions as they enter the archive. The all-sky showcase is a comparison target only.
Background
All of the above
Output
A reproducible multi-session mosaic with a complete lineage
Main issue
Most of the sessions behind the existing composite are not yet in the research archive

How a student result enters the record

  • Every new product gets a new dataset ID; existing records are never edited to match new results.
  • Its record lists what it was derived from, the code and version used, its processing level and its calibration epoch.
  • A product built from native data by documented, rerunnable code is an L8 independently reproduced product, the level the project most needs.