91 datasets for recommendation. Most are inferred from the data: a label field, the title, or the graph itself.
91 datasets.
A Delicious social bookmarking platform (tag-item network) dataset of 38,289,740 nodes and 137,240,382 edges, collected by KONECT via Delicious API. Each record contains directed: False, weighted: True, temporal: True. Massive temporal bipartite network of Delicious tags and bookmarked items — 301M edges across 38M tags and items with Unix timestamps.. Suitable for social tagging analysis, folksonomy research, tag-item recommendation.
A Delicious social bookmarking platform (user-item network) dataset of 34,611,302 nodes and 101,798,957 edges, collected by KONECT via Delicious API. Each record contains directed: False, weighted: True, temporal: True. Massive temporal bipartite network of Delicious users and bookmarked URLs — 301M edges across 35M users and items with Unix timestamps.. Suitable for social bookmarking analysis, recommendation systems, large-scale bipartite networks.
A Foursquare dataset of 22,809,624 check-ins, 114,324 users, 3,820,891 venues, collected by Yang et al. via Foursquare API (2012–2014). Each record contains check-in timestamps, venue IDs, GPS coordinates, venue categories, friendship edges. One of the largest public Foursquare datasets with both check-in histories and social networks. Suitable for location-based social network analysis, user mobility, POI recommendation, community detection.
A YouTube dataset of 743,569 videos, 14,338,803 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 0222 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
A Amazon.com (product reviews and ratings) dataset of 3,376,972 nodes and 5,743,258 edges, collected by J. McAuley & J. Leskovec via Amazon product pages. Each record contains directed: False, weighted: True, temporal: True. Large bipartite user-product rating network with timestamps and 1-5 star ratings — 5.8M edges across 2.1M users and 1.2M products.. Suitable for recommender systems, rating prediction, bipartite network analysis, e-commerce research.
A Flixster (social movie network) dataset of 2,523,386 users and 7,918,801 friendship edges, collected by Javier Parra via Flixster crawl (December 2010). Each record contains user-user friendship edges and user IDs. The dataset is notable for its friendship network from Flixster, a social movie site where users shared ratings, discovered films, and found others with similar tastes. Intended for research into social network analysis, community detection, movie recommendation systems, and taste-based social network studies.
The bipartite picture-tagging network of the image bookmarking site vi.sualize.us, as distributed by KONECT (pics_ui): 2,298,816 tag assignments by 17,122 users to 495,402 pictures, forming 997,840 distinct user–picture pairs (512,524 nodes in all). An edge records that a user tagged a picture; the file has no ratings or timestamps. The network was analysed by Neubauer and Obermayer in their study of the vi.sualize.us folksonomy. Useful for folksonomy research, tag-based recommendation and the structure of large bipartite user–item networks.
A YouTube dataset of 309,314 videos, 4,444,600 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 080327 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
A YouTube dataset of 202,320 videos, 3,916,266 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 0410 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
A YouTube dataset of 186,484 videos, 2,598,034 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 080329 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
The BookCrossing user–book network in its implicit form, as distributed by KONECT (bookcrossing_full-rating): 1,149,739 interactions between 105,278 members of the BookCrossing community and 340,523 books, collected by Cai-Nicolas Ziegler in August and September 2004. An edge records that a user interacted with a book; this version carries no rating values. Useful for collaborative filtering, implicit-feedback recommendation and the structure of user–item networks.
A Last.fm music streaming and social platform dataset of 1,084,620 nodes and 4,413,416 edges, collected by KONECT via Last.fm API. Each record contains directed: False, weighted: True, temporal: True. Temporal bipartite network of Last.fm users and individual songs they listened to, with Unix timestamps per play event.. Suitable for music recommendation, temporal bipartite networks, song preference analysis.
A YouTube dataset of 155,060 videos, 2,171,719 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 080404 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
A YouTube dataset of 117,752 videos, 2,254,098 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 0313 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
A Last.fm dataset of 976741 users, 3151283 friendships, 10.9M annotations, collected by Kompan and Bielikova, first half of 2009. Each record contains user friendship graph, user-item-tag annotations, group membership, neighbourhood scores. Last.fm social network with 99k users, 10.9M music annotations, 281k tags, and 66k groups from H1 2009.. Suitable for Social influence analysis, music recommendation, tag-based community detection..
A YouTube dataset of 154,228 videos, 2,025,178 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 080402 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
A YouTube dataset of 103,162 videos, 1,993,886 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 0329 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
A YouTube dataset of 133,611 videos, 1,735,389 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 080422 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
A YouTube dataset of 91,638 videos, 1,784,291 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 0418 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
A YouTube dataset of 88,823 videos, 1,733,223 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 0422 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
A YouTube dataset of 84,512 videos, 1,646,352 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 0430 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
A YouTube dataset of 128,359 videos, 1,683,286 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 080406 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
A YouTube dataset of 123,667 videos, 1,722,879 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 080412 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
A YouTube dataset of 76,805 videos, 1,495,423 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 0424 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
A YouTube dataset of 83,135 videos, 1,607,219 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 0413 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
A CiteULike academic social bookmarking platform dataset of 754,484 nodes and 842,426 edges, collected by KONECT via CiteULike API. Each record contains directed: False, weighted: True, temporal: True. Temporal bipartite network of CiteULike users and academic papers they saved, with Unix timestamps.. Suitable for academic recommendation, social bookmarking analysis, collaborative filtering.
A YouTube dataset of 110,122 videos, 1,392,317 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 080424 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
A YouTube dataset of 102,847 videos, 1,325,498 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 080430 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
A YouTube dataset of 66,482 videos, 1,297,320 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 0511 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
A YouTube dataset of 67,866 videos, 1,323,192 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 0509 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
A YouTube dataset of 101,151 videos, 1,368,584 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 080416 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
A YouTube dataset of 67,280 videos, 1,315,004 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 0513 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
A YouTube dataset of 62,783 videos, 1,226,091 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 0426 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
A YouTube dataset of 71,077 videos, 1,330,130 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 0314 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
A YouTube dataset of 76,128 videos, 1,458,754 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 0309 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
A YouTube dataset of 62,101 videos, 1,212,590 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 0505 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
A YouTube dataset of 96,553 videos, 1,235,773 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 080502 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
A YouTube dataset of 91,149 videos, 1,146,505 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 080512 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
A YouTube dataset of 90,231 videos, 1,123,656 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 080518 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
A YouTube dataset of 88,177 videos, 1,178,568 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 080516 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
A YouTube dataset of 85,008 videos, 1,111,027 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 080510 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
A YouTube dataset of 86,605 videos, 1,173,848 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 080414 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
A YouTube dataset of 84,052 videos, 1,080,298 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 080508 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
A YouTube dataset of 81,090 videos, 1,093,928 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 080524 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
A YouTube dataset of 83,858 videos, 1,093,943 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 080603 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
A YouTube dataset of 68,260 videos, 1,310,029 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 0303 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
A YouTube dataset of 65,652 videos, 1,260,649 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 0305 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
A YouTube dataset of 63,556 videos, 1,170,706 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 0318 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
A YouTube dataset of 75,710 videos, 1,019,605 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 080605 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
A YouTube dataset of 62,216 videos, 1,141,901 related-video edges, collected by Cheng et al. via YouTube API (2007-2008). Each record contains video metadata (ID, uploader, age, category, length, views, rate, ratings, comments), related video IDs. Early YouTube snapshot 0325 capturing the video graph and metadata from the platform's rapid growth period in 2007-2008. Suitable for social network analysis, video recommendation, information diffusion, community detection.
Delicious
Foursquare
YouTube
Amazon
Flixster
BookCrossing
Last.fm
CiteULike