{"meta":{"query_hash":"7ab68898f731","filters":{"venue":"EPJ Data Science"},"cohort_total":14,"direct_labels_cover":0,"predictions_cover":14,"exported":14,"export_cap":100000,"truncated":false,"label_status":"direct model label, unvalidated","prediction_status":"machine_predicted_unvalidated (Codex and Gemma teacher distillation)","score_status":"score_only:v0-immature-baseline","snapshot":{"source":"OpenAlex, pinned release, all 482 partitions","release":"2026-06-24","frame_built":"2026-07-12"},"permalink":"https://metacan.xera.ac/q/7ab68898f731","api":"https://metacan.xera.ac/api/v1/cohort?venue=EPJ+Data+Science"},"results":[{"id":"W2096119600","doi":"10.1140/epjds/s13688-014-0024-y","title":"Inequality and cumulative advantage in science careers: a case study of high-impact journals","year":2014,"lang":"en","type":"article","venue":"EPJ Data Science","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":42,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"Social Sciences and Humanities Research Council of Canada","keywords":"Citation; Inequality; Publishing; Citation impact; Productivity; Publication; Gini coefficient; Competition (biology); Set (abstract data type); Econometrics; Economic inequality; Sociology; Economics; Computer science; Mathematics; Political science; Library science; Economic growth; Law; Biology","score_opus":0.7717071333746156,"score_gpt":0.6827132583531631,"score_spread":0.0889938750214525,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2096119600","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":"incentives","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":"incentives","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.9975848,0.0002902056,0.000986309,0.00023515173,0.0000031529528,0.000012779047,0.00013723371,0.0000050267654,0.0007452737],"genre_scores_gemma":[0.9987128,0.00014248656,0.0007695048,0.000022242208,0.000017859813,0.000011442366,0.0001690881,0.0000052665882,0.00014925715],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.9939674,0.0034271765,0.00038938405,0.0005326743,0.0009772278,0.0007061258],"domain_scores_gemma":[0.8924592,0.08647076,0.0115539655,0.0039093886,0.0033554365,0.0022511915],"candidate_categories":["metaresearch","bibliometrics"],"consensus_categories":[],"category_scores_codex":[0.008969846,0.0002916277,0.00086548645,0.005907256,0.0022343593,0.002817765,0.0012052814,0.0013037372,0.0015689057],"category_scores_gemma":[0.0458582,0.00024297625,0.0009889078,0.012355514,0.0021178378,0.0028311696,0.002328971,0.0012000381,0.00024433367],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00037159203,0.0004383492,0.95600444,0.00018083147,0.00024457957,0.0037424923,0.009654868,0.005707019,0.000719998,0.007444176,0.0006373206,0.014854421],"study_design_scores_gemma":[0.00007895889,0.00026959178,0.92721176,0.00013411534,0.00017258317,0.0035115273,0.01726845,0.03243839,0.0014577233,0.013237315,0.0041314256,0.000088143745],"about_ca_topic_score_codex":0.012054438,"about_ca_topic_score_gemma":0.011021999,"teacher_disagreement_score":0.99409276,"about_ca_system_score_codex":0.0012985972,"about_ca_system_score_gemma":0.0010135889,"threshold_uncertainty_score":0.04743761},"labels":[],"label_agreement":null},{"id":"W2110349231","doi":"10.1140/epjds/s13688-015-0054-0","title":"Enhancing disease surveillance with novel data streams: challenges and opportunities","year":2015,"lang":"en","type":"article","venue":"EPJ Data Science","topic":"Data-Driven Disease Surveillance","field":"Medicine","cited_by":239,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McGill University","funders":"National Institute of General Medical Sciences; Economic and Social Research Council; National Institutes of Health; Santa Fe Institute","keywords":"Computer science; Data science; Public health surveillance; Social media; Public health; Key (lock); Big data; Management science; Computer security; Data mining; World Wide Web; Engineering; Medicine","score_opus":0.2809649834510432,"score_gpt":0.349388351430047,"score_spread":0.06842336797900378,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2110349231","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0668778,0.013876489,0.7884562,0.10727621,0.0013705023,0.0019784675,0.007350435,0.0016927755,0.011121134],"genre_scores_gemma":[0.2930494,0.0075149876,0.6875908,0.004395901,0.0015117766,0.0010875073,0.003917546,0.00015134894,0.00078073144],"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.93948376,0.045086794,0.003627999,0.0033726152,0.007933284,0.00049563334],"domain_scores_gemma":[0.6250343,0.30760282,0.013093391,0.029680183,0.021987233,0.0026020973],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.10886962,0.0013149631,0.0024273782,0.0056494037,0.0012174887,0.011348685,0.0034639323,0.0032461933,0.0009160823],"category_scores_gemma":[0.2281831,0.0012549123,0.0015058466,0.007335958,0.0037419011,0.020659273,0.0065264865,0.005443776,0.0005807038],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00054431136,0.00084771554,0.16605537,0.0047468557,0.000984099,0.00037176555,0.004456537,0.0534224,0.0040220297,0.12672073,0.01988054,0.61794764],"study_design_scores_gemma":[0.00017094609,0.0007208571,0.026302923,0.0029116194,0.00036247907,0.00087963755,0.0073955776,0.3740222,0.005796499,0.484297,0.09683477,0.00030561548],"about_ca_topic_score_codex":0.003999163,"about_ca_topic_score_gemma":0.004765354,"teacher_disagreement_score":0.10886962,"about_ca_system_score_codex":0.0024541426,"about_ca_system_score_gemma":0.005320423,"threshold_uncertainty_score":0.5757644},"labels":[],"label_agreement":null},{"id":"W2137811660","doi":"10.1140/epjds/s13688-014-0011-3","title":"Exploiting citation networks for large-scale author name disambiguation","year":2014,"lang":"en","type":"article","venue":"EPJ Data Science","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":53,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"Social Sciences and Humanities Research Council of Canada","keywords":"Metadata; Ranking (information retrieval); Citation; Similarity (geometry); Cluster analysis; Precision and recall; Bibliographic coupling; Parameterized complexity; Heuristics","score_opus":0.3376502821107572,"score_gpt":0.46811139645195754,"score_spread":0.13046111434120033,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2137811660","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.04042672,0.0014213067,0.9515778,0.00046347207,0.00010868,0.0001049281,0.00096499705,0.0029944405,0.001937679],"genre_scores_gemma":[0.50063837,0.0007564305,0.49256903,0.00015652027,0.00034836895,0.00022710193,0.0032823049,0.0003125059,0.0017092868],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99361455,0.0020786521,0.00054317113,0.0013564413,0.0021027266,0.00030439356],"domain_scores_gemma":[0.96888083,0.019506503,0.0046095774,0.0032699676,0.0032678447,0.00046532092],"candidate_categories":["bibliometrics"],"consensus_categories":[],"category_scores_codex":[0.0078755515,0.0015158103,0.0016483112,0.015674299,0.002226655,0.0040662563,0.0027257032,0.0026398997,0.001089376],"category_scores_gemma":[0.05275797,0.0006727188,0.001279095,0.0132458685,0.0011831394,0.0052356864,0.0030919178,0.001735655,0.001595353],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00031551876,0.00028793685,0.047761444,0.00048703776,0.0005465207,0.00038487223,0.0005664773,0.4262884,0.007856273,0.03586206,0.008151426,0.471492],"study_design_scores_gemma":[0.000015483589,0.000028508814,0.0027204393,0.00003833892,0.00003753697,0.00014523952,0.000073796604,0.9489338,0.0044947946,0.039700527,0.0037730297,0.000038526727],"about_ca_topic_score_codex":0.0047987476,"about_ca_topic_score_gemma":0.006249134,"teacher_disagreement_score":0.9843257,"about_ca_system_score_codex":0.0016928713,"about_ca_system_score_gemma":0.0025034763,"threshold_uncertainty_score":0.041650355},"labels":[],"label_agreement":null},{"id":"W2626890125","doi":"10.1140/epjds/s13688-017-0108-6","title":"Inferring social influence in transport mode choice using mobile phone data","year":2017,"lang":"en","type":"article","venue":"EPJ Data Science","topic":"Human Mobility and Location-Based Analysis","field":"Social Sciences","cited_by":72,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Calgary","funders":"Thailand Research Fund","keywords":"Interpersonal ties; Mode (computer interface); Public transport; Mobile phone; Mode choice; Phone; Social network (sociolinguistics); Computer science; Social psychology; Psychology; Telecommunications; Transport engineering; Engineering; Social media; Human–computer interaction; World Wide Web","score_opus":0.19593295680681566,"score_gpt":0.47651314161055247,"score_spread":0.2805801848037368,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2626890125","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.9930281,0.00008359454,0.0030650517,0.00009256135,0.000007557797,0.000035486333,0.002737113,0.000031892607,0.0009185579],"genre_scores_gemma":[0.99605286,0.000049892074,0.0016405039,0.00000865268,0.000009867958,0.000028211722,0.0019189487,0.0000031155646,0.00028793715],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.99909246,0.0004325925,0.000055636716,0.00017985427,0.00015147575,0.00008793446],"domain_scores_gemma":[0.9953204,0.0031673212,0.00062514987,0.00031024258,0.00035904438,0.00021780735],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001192368,0.00032417677,0.00041652427,0.0024543505,0.00040302466,0.0007290144,0.00031962473,0.00046077982,0.0014514413],"category_scores_gemma":[0.007872221,0.00018429136,0.00069666345,0.0025200562,0.00021618056,0.00069586944,0.0005548894,0.00049648224,0.0004548409],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00015664399,0.0001607453,0.97599775,0.000048284888,0.00018905921,0.00018033259,0.0005770623,0.006969525,0.0009584797,0.0007658297,0.0007085783,0.013287752],"study_design_scores_gemma":[0.00001761549,0.00018032671,0.8503845,0.000037543887,0.00019698849,0.00017073641,0.0015904955,0.14176537,0.0012615167,0.001295326,0.0030684613,0.000031187436],"about_ca_topic_score_codex":0.027304523,"about_ca_topic_score_gemma":0.030538702,"teacher_disagreement_score":0.027304523,"about_ca_system_score_codex":0.0005636466,"about_ca_system_score_gemma":0.00043548565,"threshold_uncertainty_score":0.05429119},"labels":[],"label_agreement":null},{"id":"W2803567385","doi":"10.1140/epjds/s13688-018-0166-4","title":"Prestige drives epistemic inequality in the diffusion of scientific ideas","year":2018,"lang":"en","type":"article","venue":"EPJ Data Science","topic":"Complex Network Analysis Techniques","field":"Physics and Astronomy","cited_by":88,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"","funders":"National Science Foundation","keywords":"Prestige; Competition (biology); Quality (philosophy); Citation; Set (abstract data type); Sociology; Mechanism (biology); Work (physics); Computer science; Inequality; Epistemology; Positive economics; Public relations; Political science; Economics; Mathematics; Law; Physics","score_opus":0.053099800836707976,"score_gpt":0.3519476272919705,"score_spread":0.29884782645526253,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2803567385","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":"evaluation","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":"evaluation","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.9877815,0.00020031235,0.0038312406,0.00068324036,0.0000070473843,0.000015275768,0.0000707375,0.000022491318,0.0073881885],"genre_scores_gemma":[0.9994172,0.000060354778,0.00021478528,0.00001686021,0.000008026678,0.0000041125327,0.00001442756,0.0000022184645,0.00026211297],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.9987098,0.000490877,0.00006646777,0.00020833443,0.00016256404,0.00036206807],"domain_scores_gemma":[0.9726616,0.019411555,0.0047570528,0.0011821315,0.00072911876,0.001258511],"candidate_categories":["metaresearch","bibliometrics"],"consensus_categories":[],"category_scores_codex":[0.0030943851,0.00022710442,0.0006002935,0.0022185831,0.0013936795,0.004427621,0.00066606945,0.0012445018,0.0064419894],"category_scores_gemma":[0.03614406,0.00033766334,0.0005332804,0.0017063235,0.0026793987,0.003911404,0.002629612,0.0009930033,0.00038554042],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00083354005,0.0005527612,0.6442666,0.00038646735,0.00035510908,0.0014869927,0.01290465,0.05447637,0.008293945,0.20830391,0.0021257368,0.06601392],"study_design_scores_gemma":[0.00025123978,0.00029711562,0.5109641,0.0001496417,0.00033203163,0.00064729044,0.010341536,0.17428684,0.0023596257,0.29496515,0.005251009,0.00015445029],"about_ca_topic_score_codex":0.011462166,"about_ca_topic_score_gemma":0.00847157,"teacher_disagreement_score":0.9977814,"about_ca_system_score_codex":0.0017108319,"about_ca_system_score_gemma":0.0009311607,"threshold_uncertainty_score":0.02279085},"labels":[],"label_agreement":null},{"id":"W3120906435","doi":"10.1140/epjds/s13688-020-00254-7","title":"Challenges when identifying migration from geo-located Twitter data","year":2021,"lang":"en","type":"article","venue":"EPJ Data Science","topic":"Human Mobility and Location-Based Analysis","field":"Social Sciences","cited_by":17,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McGill University","funders":"McGill University","keywords":"Geolocation; Suspect; Data science; Computer science; Set (abstract data type); Social media; Reliability (semiconductor); TRACE (psycholinguistics); Data set; Internet privacy; World Wide Web; Sociology; Artificial intelligence","score_opus":0.3333258322916927,"score_gpt":0.41532751569389453,"score_spread":0.08200168340220182,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3120906435","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.6075557,0.0042941733,0.31568143,0.018056013,0.0012720679,0.0012442636,0.028796062,0.0013529896,0.02174738],"genre_scores_gemma":[0.8775722,0.0009341434,0.10652688,0.0012235511,0.00038276377,0.0007488357,0.010233717,0.00020699642,0.0021708675],"study_design_codex":"observational","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9564514,0.030591136,0.003665089,0.002937572,0.00568354,0.0006712339],"domain_scores_gemma":[0.7519321,0.18491367,0.018398507,0.026339032,0.017412158,0.0010045504],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.034106944,0.0004579969,0.00081655325,0.00606316,0.0012997283,0.0044134143,0.0022204078,0.0016604096,0.0018547465],"category_scores_gemma":[0.2265846,0.00070462015,0.0005807636,0.009642182,0.0016852372,0.004478403,0.003548816,0.0019710474,0.002071487],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00069233665,0.00021660923,0.61474085,0.0026994182,0.00071329845,0.0010603494,0.021680024,0.0190469,0.0043118256,0.03155984,0.028040798,0.2752378],"study_design_scores_gemma":[0.000098916586,0.00029431537,0.41722587,0.0046455655,0.00032542995,0.0024761797,0.057806768,0.1280169,0.012473027,0.13306166,0.2432536,0.0003217855],"about_ca_topic_score_codex":0.010091821,"about_ca_topic_score_gemma":0.007952137,"teacher_disagreement_score":0.034106944,"about_ca_system_score_codex":0.0010764325,"about_ca_system_score_gemma":0.001477834,"threshold_uncertainty_score":0.18037689},"labels":[],"label_agreement":null},{"id":"W3165666381","doi":"10.1140/epjds/s13688-021-00281-y","title":"Inferring psychological traits from spending categories and dynamic consumption patterns","year":2021,"lang":"en","type":"article","venue":"EPJ Data Science","topic":"Digital Marketing and Social Media","field":"Social Sciences","cited_by":27,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of British Columbia","funders":"","keywords":"Extraversion and introversion; Neuroticism; Burstiness; Big Five personality traits; Psychology; Database transaction; Consumption (sociology); Trait; Personality; Social psychology; Cognitive psychology; Computer science; Computer security","score_opus":0.10829082900540199,"score_gpt":0.40998729853231647,"score_spread":0.3016964695269145,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3165666381","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.99353784,0.0000638091,0.0033625485,0.00004152363,0.000004199878,0.000011828525,0.0023864605,0.0000661623,0.00052563025],"genre_scores_gemma":[0.9962171,0.000023894023,0.0018558156,0.0000059715976,0.0000046772207,0.000012905652,0.0017614543,0.000003648603,0.00011457204],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.9995696,0.00015398322,0.000042084706,0.00010194163,0.00006698125,0.000065450666],"domain_scores_gemma":[0.995688,0.0024682553,0.00085851346,0.00041367917,0.00036477856,0.00020670796],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00076376414,0.00050345244,0.00039031426,0.0028971487,0.00019504732,0.0010646563,0.00026439913,0.00045376626,0.0018963479],"category_scores_gemma":[0.004710339,0.0001595117,0.0005621253,0.0022849338,0.00019077188,0.0005914932,0.00049760414,0.0004967859,0.00061314076],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00011391824,0.00007463947,0.980648,0.000027815482,0.00010282738,0.000042802614,0.0001289689,0.0021960118,0.00091295043,0.000094499104,0.0003511895,0.015306405],"study_design_scores_gemma":[0.000003199583,0.00006221971,0.95917785,0.000016577997,0.000032002677,0.000096842334,0.0003572359,0.03880448,0.00055362243,0.0005130415,0.00036835545,0.000014657499],"about_ca_topic_score_codex":0.0035044674,"about_ca_topic_score_gemma":0.0041184416,"teacher_disagreement_score":0.0035044674,"about_ca_system_score_codex":0.00021535048,"about_ca_system_score_gemma":0.0001476818,"threshold_uncertainty_score":0.0069681406},"labels":[],"label_agreement":null},{"id":"W4291178448","doi":"10.1140/epjds/s13688-022-00357-3","title":"Reaching the bubble may not be enough: news media role in online political polarization","year":2022,"lang":"en","type":"article","venue":"EPJ Data Science","topic":"Social Media and Politics","field":"Social Sciences","cited_by":11,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"University of Toronto","funders":"Connaught Fund; Conselho Nacional de Desenvolvimento Científico e Tecnológico; Coordenação de Aperfeiçoamento de Pessoal de Nível Superior; Fundação de Amparo à Pesquisa do Estado de São Paulo","keywords":"Politics; Polarization (electrochemistry); Presidential system; Centrality; Political science; Presidential election; Social media; Biology and political orientation; News media; Metric (unit); Computer science; Public relations; Business; Mathematics; Law; Statistics; Marketing","score_opus":0.10803353632876328,"score_gpt":0.38813913991476234,"score_spread":0.28010560358599906,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4291178448","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.9841978,0.00038394748,0.00143312,0.0017838597,0.000040170733,0.000026419888,0.00034239294,0.000017343713,0.011775064],"genre_scores_gemma":[0.99910116,0.00008861392,0.00025711418,0.00007671287,0.000028423123,0.000009122221,0.00007423641,0.0000061956684,0.0003584209],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.9985689,0.0007561766,0.00007111447,0.00017518908,0.00023199535,0.00019660781],"domain_scores_gemma":[0.972748,0.019674592,0.003814874,0.0007195074,0.0016428683,0.0014002526],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0021699923,0.00015959797,0.00024307544,0.0012746474,0.001351386,0.0037662887,0.0002349171,0.00047516625,0.0039361063],"category_scores_gemma":[0.02270984,0.00014827498,0.00019495157,0.0018074878,0.0009959617,0.002524669,0.0013959263,0.0010037238,0.00038402143],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0004074583,0.0002048704,0.9105926,0.0002285065,0.000109633926,0.00040087587,0.021999253,0.00040046606,0.0027122432,0.011570113,0.003049556,0.048324447],"study_design_scores_gemma":[0.000027166332,0.00012453171,0.897013,0.00027259174,0.0001615555,0.00046859644,0.06589654,0.006740036,0.00222849,0.009416456,0.017594194,0.000056779067],"about_ca_topic_score_codex":0.007562196,"about_ca_topic_score_gemma":0.011613733,"teacher_disagreement_score":0.007562196,"about_ca_system_score_codex":0.0008161709,"about_ca_system_score_gemma":0.00070369977,"threshold_uncertainty_score":0.0150363445},"labels":[],"label_agreement":null},{"id":"W4367308548","doi":"10.1140/epjds/s13688-023-00378-6","title":"Differences in collaboration structures and impact among prominent researchers in Europe and North America","year":2023,"lang":"en","type":"article","venue":"EPJ Data Science","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":38,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Public Health Ontario; University of Toronto","funders":"Agència de Gestió d'Ajuts Universitaris i de Recerca; Ministerio de Ciencia e Innovación; Ministerio de Economía y Competitividad; Generalitat de Catalunya","keywords":"Scope (computer science); Politics; Political science; Quality (philosophy); Network structure; Regional science; Public relations; Sociology; Computer science; Epistemology","score_opus":0.7026761929763207,"score_gpt":0.5976172621589622,"score_spread":0.10505893081735851,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4367308548","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":"evaluation","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":"evaluation","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.99292713,0.0005990235,0.00017075392,0.0002725381,0.0000052340865,0.0000063772527,0.00020925756,0.0000049562755,0.0058047553],"genre_scores_gemma":[0.99925596,0.00021410322,0.00011991023,0.000030000067,0.0000066322564,0.0000043522873,0.00015500751,0.0000017252307,0.00021232221],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.99775535,0.00077457575,0.00023087904,0.00038651229,0.000508493,0.00034424372],"domain_scores_gemma":[0.9788265,0.0075455415,0.008129787,0.0007834801,0.002876696,0.00183794],"candidate_categories":["metaresearch","bibliometrics"],"consensus_categories":[],"category_scores_codex":[0.0028183134,0.000107173844,0.00037975563,0.005764041,0.00081415835,0.0024108675,0.00027393288,0.00038503922,0.0027069282],"category_scores_gemma":[0.013081517,0.00012375858,0.00020935904,0.008111051,0.0007482179,0.0013965792,0.001954768,0.00023245998,0.00021076345],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00017549234,0.000066048364,0.9687553,0.00011107701,0.00027613033,0.00018121958,0.0053035356,0.00050611666,0.001314765,0.003010078,0.00077727233,0.019522991],"study_design_scores_gemma":[0.000010673566,0.000017637421,0.9906619,0.00006150466,0.000046646368,0.00008152078,0.0060068453,0.00024078327,0.00021815031,0.0008969467,0.001749471,0.000007889506],"about_ca_topic_score_codex":0.007006399,"about_ca_topic_score_gemma":0.011388762,"teacher_disagreement_score":0.9971817,"about_ca_system_score_codex":0.0009406579,"about_ca_system_score_gemma":0.0010853182,"threshold_uncertainty_score":0.014904857},"labels":[],"label_agreement":null},{"id":"W4379794872","doi":"10.1140/epjds/s13688-023-00394-6","title":"Do poverty and wealth look the same the world over? A comparative study of 12 cities from five high-income countries using street images","year":2023,"lang":"en","type":"article","venue":"EPJ Data Science","topic":"Urban Green Space and Health","field":"Environmental Science","cited_by":9,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"University of British Columbia","funders":"Medical Research Council; Wellcome Trust","keywords":"Atlanta; Socioeconomic status; Poverty; Inequality; Geography; Urbanization; Disadvantage; Urban studies; Metropolitan area; Regional science; Economic growth; Economic geography; Sociology; Political science; Demography; Population; Economics","score_opus":0.063362490875996,"score_gpt":0.33576062768050635,"score_spread":0.27239813680451036,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4379794872","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.99854904,0.000100204794,0.000080256694,0.00007161,0.0000028593395,0.0000063297744,0.00060122105,0.0000042172146,0.0005842868],"genre_scores_gemma":[0.99820423,0.00016922373,0.00030180716,0.000030736217,0.000004298898,0.00001152672,0.001032693,0.000011338193,0.00023411612],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.99962723,0.00011498572,0.000026223268,0.00007113738,0.00007401944,0.000086431835],"domain_scores_gemma":[0.99889815,0.0002984756,0.00031615296,0.00012844293,0.00023349552,0.00012519599],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0004366502,0.00025609529,0.00033484696,0.0024621906,0.000755672,0.0013609643,0.00024461685,0.00040731367,0.0010669958],"category_scores_gemma":[0.0023033647,0.00024379032,0.0004394425,0.003719476,0.0011536288,0.0013837287,0.0010149721,0.00042332488,0.00019522743],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00039798455,0.0001639363,0.9567832,0.0001489307,0.00028804003,0.0006847353,0.012009724,0.0011391791,0.00113756,0.0008008911,0.0034340266,0.023011839],"study_design_scores_gemma":[0.000005920321,0.000037915157,0.98239094,0.000040311235,0.000044303517,0.00015293542,0.014524217,0.00067878247,0.0001841266,0.00013531224,0.0017894752,0.000015777814],"about_ca_topic_score_codex":0.053545617,"about_ca_topic_score_gemma":0.12074764,"teacher_disagreement_score":0.053545617,"about_ca_system_score_codex":0.0006921006,"about_ca_system_score_gemma":0.0002904919,"threshold_uncertainty_score":0.1064679},"labels":[],"label_agreement":null},{"id":"W4390747461","doi":"10.1140/epjds/s13688-023-00432-3","title":"What relational event models can reveal: Commentary on Thomas Grund’s “Dynamics of Denunciation: The Limits of a Scandal”","year":2024,"lang":"en","type":"article","venue":"EPJ Data Science","topic":"Opinion Dynamics and Social Influence","field":"Physics and Astronomy","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"University of Alberta","funders":"","keywords":"Denunciation; Sociology; Event (particle physics); Context (archaeology); Scholarship; Epistemology; Dynamics (music); Politics; Value (mathematics); Computer science; Law; Political science; History; Philosophy","score_opus":0.057724232955457526,"score_gpt":0.33596203904594923,"score_spread":0.2782378060904917,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4390747461","genre_codex":"commentary","genre_gemma":"commentary","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"commentary","genre_consensus":"commentary","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.00028570465,0.0035709084,0.00077706063,0.9866377,0.007158315,0.0000031035493,0.000050319133,0.000014978364,0.0015019053],"genre_scores_gemma":[0.11371084,0.0089908745,0.0022561436,0.80719876,0.061891634,0.000119463235,0.000086615626,0.0002996212,0.005446035],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9578535,0.025648914,0.0020058283,0.005497927,0.007149832,0.0018440818],"domain_scores_gemma":[0.66897035,0.30791628,0.0040737903,0.004445097,0.011913931,0.0026804975],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.044281486,0.0015588365,0.002621365,0.0037737119,0.012813798,0.016514726,0.008230561,0.04783392,0.004441273],"category_scores_gemma":[0.19947918,0.0013787833,0.0022477403,0.0039805155,0.05250604,0.02826061,0.008228152,0.061869606,0.0021066833],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00006941756,0.000009406865,0.00016073493,0.00033870255,0.000041240208,0.00055449957,0.021084521,0.0008145856,0.00009451381,0.29355627,0.6768147,0.0064614625],"study_design_scores_gemma":[0.000039059516,0.000018250174,0.00028002713,0.002055532,0.000040396517,0.0004042579,0.0115041835,0.0025749195,0.00030385616,0.39207384,0.5904954,0.00021031911],"about_ca_topic_score_codex":0.06599886,"about_ca_topic_score_gemma":0.06192813,"teacher_disagreement_score":0.06599886,"about_ca_system_score_codex":0.02242899,"about_ca_system_score_gemma":0.0135862725,"threshold_uncertainty_score":0.23418564},"labels":[],"label_agreement":null},{"id":"W4409271193","doi":"10.1140/epjds/s13688-025-00549-7","title":"Twitter-MusicPD: melody of minds - navigating user-level data on multiple mental health disorders and music preferences","year":2025,"lang":"en","type":"article","venue":"EPJ Data Science","topic":"Mental Health via Writing","field":"Psychology","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Guelph; Toronto Metropolitan University; McMaster University","funders":"","keywords":"Mental health; Psychology; Computer science; Psychiatry","score_opus":0.23042946996787192,"score_gpt":0.4631044083909127,"score_spread":0.2326749384230408,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4409271193","genre_codex":"dataset","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.017426083,0.00023388273,0.002123133,0.0005255626,0.00019425263,0.00016910004,0.9717808,0.0028276425,0.004719655],"genre_scores_gemma":[0.03252149,0.00023277917,0.006391738,0.00022042978,0.00013168315,0.0005552671,0.9560541,0.00027532683,0.0036172918],"study_design_codex":"not_applicable","study_design_gemma":"observational","domain_scores_codex":[0.99933225,0.00014167927,0.000103074744,0.00017573364,0.00016825632,0.00007909424],"domain_scores_gemma":[0.99800056,0.0006531042,0.0003341673,0.00044902973,0.00028121276,0.0002818491],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00059876294,0.0009851289,0.00051087065,0.0032122643,0.0005946692,0.0012181777,0.00076935126,0.00089914486,0.015412246],"category_scores_gemma":[0.0051860115,0.00024704318,0.00059898355,0.002889625,0.0002920456,0.0015050224,0.0024126847,0.0008076206,0.017421601],"study_design_candidate":"observational","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00067710975,0.0002070272,0.06437675,0.002010299,0.00015904919,0.0004039973,0.0012025712,0.0013214451,0.0059726783,0.0019827217,0.8343321,0.08735423],"study_design_scores_gemma":[0.00023418678,0.00030099959,0.200662,0.00046788674,0.00013130062,0.0007075566,0.0022613066,0.013828523,0.007127722,0.004892166,0.76919234,0.00019397202],"about_ca_topic_score_codex":0.0060199285,"about_ca_topic_score_gemma":0.014771196,"teacher_disagreement_score":0.015412246,"about_ca_system_score_codex":0.00043930302,"about_ca_system_score_gemma":0.0006417576,"threshold_uncertainty_score":0.05155909},"labels":[],"label_agreement":null},{"id":"W4412385652","doi":"10.1140/epjds/s13688-025-00561-x","title":"Solana’s transaction network: analysis, insights, and comparison","year":2025,"lang":"en","type":"article","venue":"EPJ Data Science","topic":"Digital Platforms and Economics","field":"Business, Management and Accounting","cited_by":4,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Alberta","funders":"","keywords":"Database transaction; Computer science; Network analysis; Econometrics; Statistical physics; Artificial intelligence; Economics; Physics; Engineering; Electrical engineering","score_opus":0.03624091006558341,"score_gpt":0.26430469727780953,"score_spread":0.22806378721222612,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4412385652","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.8038014,0.0030242593,0.16054098,0.0020105883,0.000055516975,0.00015160088,0.0016738299,0.00036589074,0.02837602],"genre_scores_gemma":[0.98031735,0.0011686479,0.015340261,0.000045842415,0.00002034885,0.0000469671,0.00092136196,0.00003767447,0.002101562],"study_design_codex":"simulation_or_modeling","study_design_gemma":"observational","domain_scores_codex":[0.9995789,0.00015435486,0.000015539696,0.00008088183,0.000112879,0.00005750486],"domain_scores_gemma":[0.9969387,0.0019013428,0.00034893275,0.00029558488,0.00037397374,0.00014147803],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00117365,0.00034839378,0.000390862,0.0029686773,0.00054916844,0.0012992165,0.00065314915,0.0004891551,0.0023668918],"category_scores_gemma":[0.004461267,0.00015724543,0.0003801829,0.0028552006,0.00066599424,0.0030422835,0.0007110483,0.00064504094,0.0002989045],"study_design_candidate":"observational","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0004782574,0.0002281709,0.052245706,0.00032458492,0.00011905835,0.00062556134,0.0010329603,0.49231207,0.0051242956,0.3429782,0.0056085843,0.09892258],"study_design_scores_gemma":[0.000009718255,0.000056722012,0.006080561,0.000035543693,0.000017075403,0.00026320654,0.00043117176,0.9117959,0.001070614,0.07370385,0.0065162797,0.00001937356],"about_ca_topic_score_codex":0.0044899657,"about_ca_topic_score_gemma":0.0049623367,"teacher_disagreement_score":0.0044899657,"about_ca_system_score_codex":0.0012706891,"about_ca_system_score_gemma":0.0006578091,"threshold_uncertainty_score":0.009219587},"labels":[],"label_agreement":null},{"id":"W4413745650","doi":"10.1140/epjds/s13688-025-00582-6","title":"AGECovP: identifying ageism and analyzing COVID-19 discourse on older adults in YouTube","year":2025,"lang":"en","type":"article","venue":"EPJ Data Science","topic":"Aging and Gerontology Research","field":"Psychology","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Western University; Toronto Metropolitan University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Coronavirus disease 2019 (COVID-19); 2019-20 coronavirus outbreak; Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2); Computer science; Psychology; Sociology; Medicine; Virology","score_opus":0.11659928807268619,"score_gpt":0.5068965799335734,"score_spread":0.39029729186088724,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4413745650","genre_codex":"dataset","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.20136248,0.0015747915,0.0037214134,0.0014631364,0.0004751327,0.0011245892,0.7598631,0.0013037788,0.029111572],"genre_scores_gemma":[0.18385941,0.0016091292,0.016105533,0.0005602306,0.00039154806,0.004040383,0.77354425,0.00042352997,0.01946593],"study_design_codex":"not_applicable","study_design_gemma":"observational","domain_scores_codex":[0.9991702,0.00023844109,0.00008758684,0.00016028096,0.00022536427,0.000118187505],"domain_scores_gemma":[0.9971898,0.0012237874,0.00039738993,0.00030592878,0.0006743569,0.00020871186],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009984933,0.00062175246,0.0003185506,0.004885939,0.0010369137,0.0012709129,0.00040964535,0.0006996251,0.007969651],"category_scores_gemma":[0.005287041,0.00013939023,0.00029336585,0.0029818278,0.00045549625,0.0022909292,0.0023517613,0.00062231295,0.004558904],"study_design_candidate":"observational","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00032855105,0.00019073268,0.09784067,0.0034895788,0.00008140822,0.0010932981,0.027066913,0.00053828827,0.0045250715,0.0049553015,0.70392114,0.15596896],"study_design_scores_gemma":[0.00005727915,0.00014935285,0.26356348,0.0013844203,0.000046689907,0.0005414812,0.031658527,0.0053483048,0.0035215393,0.0035038698,0.6900902,0.00013482384],"about_ca_topic_score_codex":0.027641946,"about_ca_topic_score_gemma":0.08946681,"teacher_disagreement_score":0.027641946,"about_ca_system_score_codex":0.0010407675,"about_ca_system_score_gemma":0.0011181388,"threshold_uncertainty_score":0.0549621},"labels":[],"label_agreement":null}]}