{"id":"W6950346504","doi":"10.5281/zenodo.7998268","title":"Successes, pain points, and lessons learned when curating data at scale: The NACJD experience","year":2023,"lang":"en","type":"article","venue":"Zenodo (CERN European Organization for Nuclear Research)","topic":"Data Analysis and Archiving","field":"Social Sciences","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Canadian Institute for Public Safety Research and Treatment","funders":"","keywords":"Workflow; Restructuring; Criminal justice; Economic Justice; Presentation (obstetrics); The Internet; Period (music); Data curation","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.1972468,0.000971317,0.001289595,0.005174175,0.01583045,0.03798104,0.009645014,0.007186963,0.006961583],"category_scores_gemma":[0.249208,0.002195276,0.001618189,0.008281281,0.02486355,0.0434199,0.03517137,0.01636656,0.003565732],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0140232,"about_ca_system_score_gemma":0.02462303,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.04556848,"about_ca_topic_score_gemma":0.06846652,"domain_scores_codex":[0.8503219,0.07800457,0.01211483,0.01815262,0.03491933,0.006486755],"domain_scores_gemma":[0.677506,0.1645509,0.007492244,0.05059489,0.05874707,0.04110884],"domain_codex":null,"domain_gemma":"reproducibility","domain_candidate":"reproducibility","domain_consensus":null,"study_design_codex":"qualitative","study_design_gemma":"observational","study_design_scores_codex":[0.0002271571,0.0005926311,0.02367371,0.002411575,0.0001059182,0.002849486,0.4331381,0.0008167397,0.005417694,0.04596105,0.2505491,0.2342568],"study_design_scores_gemma":[0.000035672,0.00009091045,0.004659607,0.001012331,0.00002216115,0.001463578,0.102587,0.0009218744,0.001409271,0.01144217,0.8761287,0.000226698],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"commentary","genre_gemma":"empirical","genre_scores_codex":[0.1517734,0.03220489,0.1030398,0.5971275,0.01151173,0.0009907903,0.003593055,0.005494562,0.09426422],"genre_scores_gemma":[0.4135633,0.02946004,0.3787115,0.1020448,0.00433254,0.001298112,0.005591776,0.009948963,0.05504893],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.8027532,"threshold_uncertainty_score":0.9899376,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1384627393882129,"score_gpt":0.3563165711805265,"score_spread":0.2178538317923136,"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."}}