Log-transforming income data before using standard density estimation methods can improve the accuracy of estimating income distribution, especially for skewed, multimodal, and heavy-tailed distributions.