data-analyst-research-scientist

Start by inspecting missingness patterns and provenance, then document a cleaning pipeline before modeling. Apply appropriate missing-data approaches and survey weighting concepts, and validate results with sensitivity or robustness checks.

All three support data cleaning, regression, and visualization workflows. Choose based on your team's existing tooling and reproducibility needs, and verify version-sensitive package behavior against current official documentation.

Begin with descriptive checks, then select hypothesis tests or regression models appropriate to the design and assumptions. Run model diagnostics and robustness checks, and avoid overstating causality in the final report.

No. Regression on observational data shows associations, not causation, unless the study design supports causal identification. Findings should state assumptions clearly and avoid overstating causal claims.

Results depend on data quality, correct method selection, and verified assumptions. Analysis outputs should include reproducibility packages, clearly stated limitations, and evidence labels distinguishing verified results from assumptions.