This dataset contains raw (unprocessed Raman spectra) of urine from de-identified human patients. Analysis of this dataset is included in our journal article, "Analysis of urine Raman spectra differences from patients with diabetes mellitus and other pathologies." The dataset includes urine Raman spectra from (1) healthy volunteers, (2) patients with chronic kidney disease and diabetes mellitus, (3) patients with chronic kidney disease and without diabetes mellitus, (4) patients with biopsy-confirmed diabetic nephropathy, (5) patients with biopsy-confirmed immune-mediated nephropathy, (5) patients with biopsy-confirmed membranous nephropathy, (6) patients with biopsy-confirmed renal neoplasm, (7) patients with other glomerular pathologies, and (8) Surine (urinalysis control). Raman spectra were obtained with the following parameters: Raman spectrometer: Agiltron PeakSeeker PRO-785 Mode: Bulk liquid scanning Wavelength: 785 nm Wavenumber range: 200-2000 cm-1 Laser power: 30 mW Spectral resolution: 8 cm-1 Laser spot size: 0.2 mm Excitation time: 30 s Raman spectral data and de-identified metadata are present in tab-separated value (tsv) files. Each urine sample is bar-code identified and linked to a disease state (or control) in the metadata tsv file. Ten (10) independent Raman scan replicates exist for each urine sample and are identified by bar-code in the spectral data tsv file. The study IRB approval and a sample blank patient consent form are also included here.