{"id":"W2611926229","doi":"","title":"Beyond the Dataset: Understanding Sociotechnical Aspects of the Knowledge Discovery Process Among Modern Data Professionals","year":2017,"lang":"en","type":"dissertation","venue":"UWSpace (University of Waterloo)","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Natural Sciences and Engineering Research Council of Canada; University of Waterloo; Simon Fraser University","keywords":"Sociotechnical system; Data science; Process (computing); Knowledge management; Computer science","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["open_science"],"consensus_categories":[],"category_scores_codex":[0.003144572,0.0002403961,0.0005227636,0.0001864311,0.001184781,0.0002340077,0.009488006,0.0002768573,0.000141079],"category_scores_gemma":[0.000580921,0.0001462603,0.0001718293,0.0002918082,0.0008242002,0.002014741,0.002275838,0.0004757846,0.00002705349],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00008555248,"about_ca_system_score_gemma":0.0003892413,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004145317,"about_ca_topic_score_gemma":0.1321588,"domain_scores_codex":[0.9966102,0.0003910758,0.0003567615,0.0007855097,0.001592213,0.0002642669],"domain_scores_gemma":[0.9938381,0.0005172441,0.001428073,0.003962664,0.0001907863,0.0000630828],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"qualitative","study_design_scores_codex":[0.0006535186,0.001149785,0.001037429,0.001654947,0.0009122986,0.00003939406,0.3143585,0.0001100805,0.0005797609,0.09240181,0.5752497,0.01185278],"study_design_scores_gemma":[0.0007056611,0.00005524851,0.01254113,0.0009042231,0.0005211087,0.000001108236,0.8018394,0.00146565,0.0004500635,0.1756271,0.005396112,0.0004932685],"study_design_candidate":"qualitative","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.921321,0.0005471912,0.001754534,0.01915648,0.002852288,0.003001556,0.02587305,0.00007060792,0.02542333],"genre_scores_gemma":[0.6975551,0.00005503296,0.00005056052,0.00003935398,0.00004796183,9.318571e-7,0.005909265,0.00001839625,0.2963234],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.5698536,"threshold_uncertainty_score":0.9958711,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1558935873536148,"score_gpt":0.3803340841839857,"score_spread":0.2244404968303708,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}