Every galaxy in the universe is in motion — not just moving away from us as space expands, but also pulled sideways by gravity toward massive concentrations of matter. This atlas maps those gravitational flows across more than a billion light-years of our cosmic neighbourhood, using real galaxy positions and velocities.

The coloured streamlines trace paths that matter follows under gravity. Every line converges toward a gravitational attractor — a supercluster massive enough to reshape the motion of thousands of galaxies. The region of space whose matter drains toward the same attractor is called a Basin of Attraction.

The Fresnel-edged shells show the five most probable basins identified by the analysis, glowing at their edges where basin boundaries meet. The Milky Way sits near the boundary between two: Laniakea/Ophiuchus (~48% probability) and Shapley (~50%). Our cosmic fate is genuinely uncertain.

A second streamline model — the 5-basin probabilistic streamlines — offers an alternative view computed using HAMLET mean-field reconstruction and HDBSCAN clustering. Flowlines are seeded within surfaces where galaxies have >50% probability of belonging to each attractor.

The Galaxy Density Fields (5 levels) show where infrared-luminous matter concentrates — the bright nodes, filaments, and walls of the cosmic web — computed directly from 20,163 galaxy positions in the 2MASS Redshift Survey (FoG-corrected, Kourkchi & Tully 2017). Five smoothing levels range from sharp cluster cores (Level 1) to the full luminosity envelope (Level 5). Reliable within ~133 Mpc.

Galaxy survey overlays (SDSS DR18, DESI DR1) extend the view far beyond the CF4 volume, revealing large-scale structure across up to 8 billion light-years. The Local Universe mode shows 20,163 nearby galaxies within ~350 million light-years, sized by intrinsic infrared brightness, with named cluster identification and group structure.

Use the Views menu at the top of the panel to jump between curated perspectives — from the full cosmic overview to close-ups of the Great Attractor, Ho'oleilana, or deep survey fields.

The velocity field was reconstructed using HAMLET (Hamiltonian Monte Carlo), a Bayesian algorithm that infers the 3D density and peculiar velocity field from galaxy redshifts across 1,000 statistical realizations. This captures uncertainty in the reconstruction.

Basins of Attraction were identified by integrating streamlines forward in the velocity field until they converge. All points whose streamlines terminate at the same attractor form one basin — a gravitational watershed, analogous to river drainage basins on Earth but in 3D, at cosmic scale.

Two streamline models are available. The primary model uses 9 basins extracted from single consistent HAMLET realizations, pre-computed using SDvision and colour-assigned by endpoint attractor. Laniakea uses a dedicated high-density model (3,162 streamlines). The 5-basin probabilistic model uses HDBSCAN clustering on the HAMLET mean field to define probabilistic Basins of Attraction (pBoA), with flowlines seeded within surfaces where galaxies have >50% membership probability.

Each CF4 galaxy is assigned to a basin by looking up its 3D position in the 128³ watershed grid. The "By basin" color mode uses these assignments, matching each galaxy to its streamline basin colour.

The Galaxy Density Field is constructed from 2MRS log-density data using Marching Cubes isosurface extraction with Gaussian blur (σ = 0.8 cells), a log-density threshold of 8.5, and Taubin smoothing (12 iterations). A hard cutoff at 10,000 km/s (~133 Mpc) ensures reliability — beyond this distance, incompleteness corrections in the 2MRS sample become too large.

Distances in the local universe are not straightforward. We observe galaxy redshifts — how fast each galaxy appears to recede — but converting that into a true distance requires care. For nearby galaxies, direct measurements (from pulsating Cepheid stars or the brightness of red giant stars) give precise distances. For the majority of galaxies, we use their recession velocity in the cosmic microwave background frame, adjusted by a cosmological correction factor (f_mod) that accounts for the geometry of expanding space.

Galaxy clusters present a different challenge: galaxies orbiting within a cluster have random motions that smear their apparent positions along the line of sight — an effect called "Fingers of God" that makes round clusters look elongated, pointing at us. We correct this using partial compression: each galaxy's radial offset from its group centre is scaled down based on the group's physical size and internal velocity spread. This removes the artificial elongation while preserving the real three-dimensional shape of clusters like Virgo and Fornax.

Galaxy brightness in the local view uses absolute K-band magnitude (M_K) — intrinsic infrared luminosity corrected for distance — so that a galaxy's apparent size reflects its true luminosity rather than its proximity.

Cosmic web — Dark matter distribution from the IllustrisTNG cosmological simulation (TNG Collaboration). Shows the filamentary large-scale structure that the real galaxies in this atlas trace.

Planck CMB — The Cosmic Microwave Background as mapped by ESA's Planck satellite — the oldest light in the universe (380,000 years after the Big Bang). The tiny temperature fluctuations visible here are the seeds from which all the structures in this atlas grew.