Cannabis-derived products (CDPs) are widely available and diverse. A classification of product and flavor types is necessary to establish a foundation for comparative research, although research aiming to classify cannabis in its variety of products, flavors, and cannabinoid compounds based on public online e-commerce data is lacking. To analyze data from a large cannabis e-commerce marketplace to identify and characterize cannabis product types, flavors, and compound types. This qualitative study was conducted in 2 phases: (1) data mining of the cannabis e-commerce website Weedmaps for product listings in the US between September 1 and November 30, 2023 and (2) grouping CDPs into product, flavor, and cannabinoid compound categories. Cannabis product listings. Product listings and descriptions were extracted from the platform. Coding was performed for specific product characteristics, routes of administration (ROAs), and characterization of flavors. A total of 573 854 unique US CDP sales listings from the platform were collected; after removing 72 842 nonconsumable items, 501 012 were analyzed. Product ROAs included multisystem (205 637 [41.04%]), respiratory (185 296 [36.98%]), digestive (98 941 [19.75%]), epidermal (9487 [1.89%]), and oral (1651 [0.33%]). Nearly half (210 575 [42.03%]) of all product listings included at least 1 flavor, with 247 762 instances of flavors. The 3 most common flavors were lemon (22 106 [8.92%]), cake (19 463 [7.86%]), and strawberry (13 961 [5.63%]). The most common cannabinoid compound type was Δ9-tetrahydrocannabinol (54 699 [63.30%]). This qualitative study categorized more than half a million CDPs for product and flavor types. Results are needed for comparative studies on product and market availability and can help in assessing concerns about appealing characteristics. The results can also inform future market surveillance efforts aimed at identifying new and emerging products as cannabis policy continues to move toward greater legalization.