outlier-detection-and-quality-assessment

Detect outliers in Excel data by loading the spreadsheet into a pandas DataFrame and computing IQR thresholds per numeric column. The 1.5×IQR rule flags anomalous values, making it suitable for datasets that do not follow normal distributions.

Assess data quality by computing skewness and kurtosis to evaluate distribution shape. These metrics characterize whether your data is symmetric, skewed, or heavy-tailed, helping you make safer preprocessing decisions before analysis.

Yes, you can generate boxplot visualizations for outlier detection across multiple numeric columns. The process produces an annotated PNG report containing boxplots and descriptive statistics to quickly review distributions and anomalies.

IQR-based outlier detection works effectively for non-normal distributions. By calculating interquartile range thresholds per column, it identifies extreme values without assuming normality, making it robust for operational metrics and skewed data.

Decide whether to cap or drop outliers by reviewing the generated per-column quality report. The report combines IQR statistics with skewness, kurtosis, and boxplot visualizations to guide your data cleaning choices based on distribution characteristics.

IQR for outlier detection may be less effective when extreme values represent valid observations rather than errors. If your data requires multivariate anomaly detection or has highly complex dependencies, IQR-based univariate rules might flag legitimate data points as outliers.