The analysis used topological data analysis to examine the geometric structure of store performance data, revealing distinct patterns and clusters that weren't visible through traditional statistical approaches. Store locations with similar characteristics were connected in a network visualization, with different metrics represented by color coding.

‍

The analysis revealed several critical insights:

• A distinct cluster of stores showing anomalously high turnover rates disproportionate to other locations with similar wage levels and market conditions
• Stores within certain geographic regions were overpaying relative to local unemployment rates while achieving suboptimal retention
• Locations where unemployment levels were low exhibited stronger correlation between wages and turnover, signaling these markets needed higher compensation to retain talent
• "Flares" in the data network identified groups of stores requiring targeted intervention

Through this analysis, the company was able to optimize its labor costs by identifying precise wage levels needed for each market while improving employee retention to the tune of hundreds of millions of dollars.