Galaxies evolve in regular, bi-directional communication with their environment. Large-scale structure informs galaxy evolution via mergers, gas accretion via the Cosmic Web, etc., and galaxies release material and energy reaching tens of kpc from their disks via feedback processes. This thesis investigates local large-scale structure and the global gas properties of feedback-hosting galaxies via spectral line observations of atomic and molecular gas in galaxies. First, we study the dark matter mass density and infall onto supercluster-scale filaments: though filamentary structure is regularly characterized in simulations, observational constraints remain sparse. A primary difficulty is the dearth of redshift-independent distance measurements for local galaxies, which are necessary for measuring peculiar velocities. We address this by presenting and calibrating a Baryonic Tully-Fisher Relation (BTFR) using the 31,000 galaxy ALFALFA sample. We discuss the role of sample demographics and data quality in setting the slope and intercept of this BTFR, which is relevant for application of this template and for interpreting the BTFR in the context of the galaxy-halo connection. We find that changes to the sample mass range (and to second order, mass sampling, gas fraction, and different stellar mass and velocity width measurements) can significantly change the slope of a BTFR. We find that these distances have average uncertainties of ~0.17 dex for the full ALFALFA sample, and that they generally agreement with flow-model distances. We then use these distances to measure peculiar velocity: first, we measure infall onto Virgo consistent with previous work. We then derive ~2,000 distances in the direction of Pisces-Perseus, increasing the number of sources used in a peculiar velocity study in that region by an order of magnitude. Using this data, we recover a marginal signature of infall onto PPS. Though this marginal signal exceeds that of off-filament comparison samples, the best-fit values for the amplitude and scatter of the on-PPS sample are unphysically large. This is likely an indication that the model does not sufficiently capture the complexities of the large scale motions within the observed sample. We apply a complimentary technique to estimate the linear mass density along PPS using APPSS redshift data and data from the 2MASS Redshift Survey (2MRS). This allows us to make linear dark matter density estimates along 80% of the PPS filament. Finally, we pivot to studying the excitation of CO lines in a sample of outflow-hosting ULIRGs in order to characterize the conditions of the molecular ISM and the reach of potential outflow-driving mechanisms. We supplement existing mid- and high-J Herschel observations with ACA observations of the 1-0, 2-1, and 3-2 transitions. We compare the conditions in and predicted CO emission of photo-dissociation regions (PDRs), which we independently estimate using fine-structure lines and the far-infared continuum luminosity. We find that the low-J line emission in about half of these sources can be explained by PDR models, and find the strongest constraints from these models in the least obscured sources. Additionally, comparing source properties and CO excitation, we find the strongest relationships between excitation and dust temperature and silicate absorption depth, which point to the role of these sources' deeply obscured nuclei in supplying energy to the observed molecular outflows.