data-analysis

Statistical analysis on a CSV file computes descriptive summaries, Pearson correlations, and rolling average trends using pandas and scipy. It extracts actionable insights by loading structured datasets directly and applying inferential statistics to identify significant relationships.

Time-series decomposition and anomaly detection on Excel data are fully supported using statsmodels and numpy. It decomposes seasonal trends via rolling averages and flags anomalous observations using IQR and z-score methods to contextualize outliers for decision-making.

Testing hypotheses and running ANOVA on structured datasets is handled by applying scipy and statsmodels to compute p-values and confidence intervals. It executes t-tests, ANOVA, and non-parametric alternatives to validate whether observed differences in your data are statistically significant.

This approach works reliably with JSON and Parquet inputs alongside CSV and Excel formats for data manipulation. It uses pandas to load and manipulate structured datasets, ensuring dependable extraction of descriptive and inferential insights across all supported file types.

Detecting outliers using z-score and IQR methods involves computing statistical thresholds across your structured dataset to flag anomalous observations. It contextualizes these outliers for decision-making by quantifying their deviation from descriptive norms using numpy.

Use non-parametric alternatives instead of t-tests when your structured datasets violate normality assumptions. It computes these alternatives alongside ANOVA and p-values using scipy, ensuring your inferential statistics remain valid without relying on strict distributional requirements.