Data Science Pipelines and Workflow Automation
NCA-ADS · 26 questions
- A county social-services case analytics team is designing a new permitting risk pipeline. What is the sound end-to-end stage order?
- A city planning zoning analytics team wants to promote a transit demand model into the shared model store. Where should a validation gate sit in the pipeline?
- A city open-data portal team wants multiple civic models to reuse the same curated predictors. How should they structure the pipeline?
- A county public-health analytics team must design a pipeline auditors can follow from source files to scored risk. What practice best supports that?
- A transit authority ridership desk discovers that a “final disposition code” written after the routing decision is still in the feature set. What should they do before retraining?
- A water-utility sensor analytics team sees heavily right-skewed fee amounts that distort RAPIDS training. Which transformation best improves the feature for modeling?
- An emergency-management fusion cell has dozens of noisy civic predictors and limited GPU time. How should they shrink the feature set while keeping strong signals?
- A housing-authority waitlist analytics model underperforms on rare permit types after launch. What feature-engineering response fits Domain 3 practice?
- A courts and clerk e-filing analytics team sees both training and validation error stay high on a filing-volume model. What adjustment fits underfitting?
- A library program attendance analytics team watches training scores soar while validation collapses. What should they do?
- A parks and recreation demand desk finds a housing-adjacent demand model is memorizing parcel IDs instead of learning transferable patterns. What should they change?
- A school-district enrollment analytics team wants a deeper RAPIDS model. When is increasing complexity justified?
- A county assessor parcel desk has too few rare positive fraud-flag samples for stable training. What augmentation approach fits?
- A 311 resident-service analytics team wants better storm-day learning for ridership-adjacent demand models. How should they enhance training data?
- A permitting and inspections bureau is augmenting training rows and notices a proposed join would attach outcomes that only exist after the inspection date. What should they do?
- A fire and EMS response-time analytics team has scarce positive high-severity labels. Which augmentation approach is appropriate?
- A public hospital district throughput desk receives new curated Parquet nightly and wants consistent model refresh. What automation fits?
- An elections operations (non-partisan) data team’s citywide RAPIDS pipeline is hitting single-GPU memory limits as voter-file hygiene tables grow. How should they scale the workflow?
- A municipal fleet telematics desk wants one automated pipeline to serve all boroughs instead of forking per district. What design choice supports that?
- A solid-waste route optimization analytics team must ensure Monday council dashboards refresh on time. What automation reliability practice fits?
- A county social-services case analytics team needs last week’s RAPIDS and Dask pipeline rerun to match when auditors ask. What should they pin?
- A city planning zoning analytics team uses multi-stage Dask GPU tasks and needs auditability between stages. What should they materialize?
- A city open-data portal runs a RAPIDS and Dask nightly pipeline and wants reproducible results when the GPU worker pool changes. Which guard best protects reproducibility?
- A county public-health analytics team wants production prediction errors to continuously improve the next training corpus. Which approach best enables that data flywheel?
- A transit authority ridership desk wants human inspector overrides on risk scores to continuously improve the deployed model. Which flywheel step is most appropriate?
- A water-utility sensor analytics team enables a data flywheel but must stop poisoned or biased new rows from silently entering training. Which safeguard fits best?