Todd Lohr and Bill Nowacki, advisors proficient in emerging and data-driven technology, work with KPMG Signals Repository, an accelerator for data scientists and next generation developers. By creating a “big data fabric” of internal and external data, organizations can find the right data and signals to enable machine learning technologies to achieve unprecedented accuracy in predictions and help drive better business decisions.
In response to COVID-19, the Signals Repository has on-boarded new disease-relevant signals that help to pinpoint such things as locations that are most at risk for the next outbreak or locations that will have a dramatic swing in demand for goods and services.
Topics addressed include:
- how machine learning evaluation of 60,000 repository signals, geocoded to each latitude and longitude in the US, can help explain what is happening at a hyper-local level
- the fascinating insights revealed when traditional published data is geotagged (e.g., sources such as Medicare, FDIC or US Census records) and converged with new “pop-up data” (e.g., COVID-19 drive-up testing sites)
- why static data doesn’t provide needed context, but refreshed data over time and geography does
- the importance of using sensors to perceive early movements, such as an initial economic downturn or the detection of an early market upturn
- how seemingly unrelated factors such as the presence of public transit, store hours, and the ratio of senior citizens can form connections that paint a vivid picture of COVID-19 outcomes.