This visualization uses a random forest classifier to analyze CRISPR-Cas9 experimental data from an Excel file, predicting whether guide and donor sequences yield correctly targeted mouse alleles. The code computes nucleotide and dinucleotide frequencies from sequence regions, then trains a 2000-tree RandomForestClassifier with balanced class weights using scikit-learn. It displays out-of-bag error rates, the top ten features ranked by importance, and confusion matrices from StratifiedKFold cross-validation, revealing which sequence features most strongly influence targeting success across multiple laboratory centers.