{"id":"W4413133192","doi":"10.3390/computers14080319","title":"Optimizing Data Pipelines for Green AI: A Comparative Analysis of Pandas, Polars, and PySpark for CO2 Emission Prediction","year":2025,"lang":"en","type":"article","venue":"Computers","topic":"Vehicle emissions and performance","field":"Engineering","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"Artificial Intelligence in Medicine (Canada)","funders":"","keywords":"Pipeline (software); Computer science; Energy consumption; Pipeline transport; Geospatial analysis; Efficient energy use; Real-time computing; Engineering; Remote sensing; Operating system; Electrical engineering","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":[],"consensus_categories":[],"category_scores_codex":[0.001406141,0.001638933,0.0006949081,0.001038786,0.0006474248,0.00129534,0.002492138,0.0005815413,0.002994706],"category_scores_gemma":[0.003901607,0.0006286033,0.0008002711,0.002162558,0.0006573588,0.004244784,0.001253723,0.001214524,0.00119294],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001399115,"about_ca_system_score_gemma":0.003000836,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.02502647,"about_ca_topic_score_gemma":0.02519232,"domain_scores_codex":[0.9991506,0.0001232232,0.0000719178,0.0002499915,0.0002657532,0.0001384224],"domain_scores_gemma":[0.9983011,0.0007811378,0.00009038166,0.0003737817,0.0003628783,0.00009085143],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.004062857,0.00108301,0.02225826,0.001396007,0.0005850507,0.0004033218,0.0003970394,0.2911947,0.03715992,0.005882079,0.02606593,0.6095119],"study_design_scores_gemma":[0.0001792671,0.0006770029,0.00634509,0.00004081804,0.0001469715,0.0001068878,0.0002898138,0.9366586,0.04241301,0.002502241,0.01058231,0.00005812462],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.832447,0.004749984,0.09159718,0.000900869,0.0003465037,0.0003184097,0.003511331,0.05099133,0.01513738],"genre_scores_gemma":[0.8534503,0.001451794,0.1313595,0.0003290012,0.00004087633,0.0001727829,0.00818592,0.001814942,0.003194949],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.02502647,"threshold_uncertainty_score":0.04976159,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03732970444766449,"score_gpt":0.3054617075539457,"score_spread":0.2681320031062813,"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."}}