Title: Hidden Value in Outliers: New Realities Across Industries

Data is not ‘science,’ but keys to unseen ‘new realities’ or ‘unveiled truths,’ as Dirac would say, that often hide behind dataset anomalies or outliers.

Outliers, especially in retail industries focusing on high-frequency food products, represent untapped profit. These ‘Outliers,’ due to misinterpretations, miscalculations, or hidden biases, lead to wastage. –These realities, perhaps akin to what Dirac referred to as ‘unveiled truths’ —a concept I prefer to call the “Dirac Eye”. Addressing these issues is often not financially viable due to the industry’s low-profit margins.

Innovators like Rick Rijkse, however, have proposed practical solutions to tackle this wastage. Their work emphasizes the business potential, an insight further illuminated when applied to high-profit industries where the value of rectifying even a single outlier can be substantial.

Consider Werfen’s presentation at the 2023 AACC Meeting & Expo on reducing sample quality errors. eg. Werfen’s GEM Premier 5000, a blood sample exposed to air may cause carbon dioxide to escape, leading to an underestimated pCO2 value and potential misdiagnosis. This demonstrates the crucial importance of identifying and correcting outliers.

It’s clear that high-profit industries are often more driven to explore these anomalies. After all, the value of a human life surpasses that of a single watermelon. But does AI perceive this discrepancy? Perhaps, even a watermelon might hold hundreds of potential new lives…

#data #ai #business #retail #science #testing

2023.7.13. 1200. xiaowen kang

References:

  1. Rijkse, R. Practical Retail Wastage Solutions

  2. Werfen. Reducing Blood Gas Testing Errors

Figures:

  1. Fig1: Retail industry outlier issues. sometimes, stem from simple issues like store clerks failing to front-face merchandise, or customer unwillingness to bow to reach.

  2. Fig2: The Henderson-Hasselbalch equation.

  3. Fig3: Air-exposed sample impacts on results. Code: https://github.com/williampolicy/code_vitro_diagnostics.git