fix issue in eda process when NA values are present

[mitokic]

This pull request includes a version bump and minor improvements to the data preprocessing steps in the exploratory data analysis (EDA) functions. The most important changes are focused on improving data quality by handling missing values before analysis.

Data preprocessing improvements:

• In the run_xreg_analysis function in R/agent_eda.R, added a step to drop rows with any NA values from the filtered input data before analysis to ensure cleaner data is used.
• In the run_all_eda_per_combo function in R/agent_eda.R, added a step to drop rows where the Target column is NA after filtering by date, improving the reliability of downstream analysis.

Version updates:

• Bumped the package version from 0.6.0.9015 to 0.6.0.9016 in both the DESCRIPTION and NEWS.md files to reflect these changes. [1] [2]

[Copilot]

Pull request overview

This pull request adds NA value handling to two EDA functions to prevent analysis failures when missing values are present. The version is bumped from 0.6.0.9015 to 0.6.0.9016 to reflect these changes.

• Added tidyr::drop_na() to run_xreg_analysis to remove rows with any NA values before external regressor correlation analysis
• Added tidyr::drop_na(Target) to run_all_eda_per_combo to remove rows with NA Target values before running EDA analyses
• Updated version number in DESCRIPTION and NEWS.md

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 3 comments.

File	Description
R/agent_eda.R	Added NA value filtering in two EDA functions: aggressive filtering (all NAs) in run_xreg_analysis and targeted filtering (Target NAs only) in run_all_eda_per_combo
NEWS.md	Updated development version number to 0.6.0.9016
DESCRIPTION	Updated package version to 0.6.0.9016

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Using tidyr::drop_na() without arguments drops rows with NA in ANY column (including Target, Combo, Date, and all regressors). This is overly aggressive for external regressor analysis.

Looking at the correlation calculation below (lines 1408-1415), NA values are already handled appropriately per-regressor with keep <- !(is.na(x) | is.na(y)). The current approach may unnecessarily drop rows where a specific regressor being analyzed has valid values, but another regressor in the dataset has NA.

Consider using tidyr::drop_na(Target) instead to only drop rows where the Target is missing, allowing the existing per-regressor NA handling logic to work as designed.

Suggested change

	tidyr::drop_na() # drop rows with any NA values
	tidyr::drop_na(Target) # drop rows where Target is NA