The virus takes hold late in the year in the coasts in the South and the Southeast. Jumping from social circle to social circle, it gradually spreads inland and to the North, until it has spread throughout the United States. Within months, it kills thousands.

The new model of the cyclic reappearance of seasonal influenza outbreaks in the U.S., published in the journal eLife, duplicates some of the active monitoring systems run by the CDC and the various state health agencies.

But the scientists from the University of Chicago Medical Center say their new analysis—which incorporates weather patterns, social connectivity networks and even Twitter data—adds a new level of detail to the annual epidemics.

“Seasonal influenza cycles initiate in the South/Southeastern U.S. and sweep the country from south to north,” they conclude. “This pattern is repeated, with some variation, each season.”

The clinical data source was the Truven Health Analytics MarketScan Commercial Claims and Encounters Database from 2003 to 2012—a dataset that includes health insurance claims of as many as 45 million people annually.

The antigenic drift was assessed by the influenza research database of the National Institute of Allergy and Infectious Diseases. Vaccination extent was looked at by electronic health records where available.

Twitter provided general human travel patterns, with geomapping location sets. It including 1.7 billion messages over more than three years.

These datasets were all cross-referenced by weather data from the years surveyed.

“The strongest predictor groups (ranked by importance) are the population’s socio-demographic properties, weather, antigenic drift of the virus, land-based travel, and auto-correlation of influenza waves,” they write.

Weather influences the beginning of the viral spread by humid and warm areas. The antigenic drift is marked by HA, NA and M1 inhibitors—and an M2 enhancing diversity of the spread. Land travel is an especial boost to the viral dissemination.

The outbreak is like a forest fire, the doctors said in their analysis. The dry kindling that starts it all is the high degree of social connectivity among communities in the South, they contend.

The researchers contend that it may even be a clearer picture than the weekly updates provided by the CDC from the fall into the spring.

“It’s a very high-resolution picture, perhaps even higher than what the Centers for Disease Control and Prevention can see, because it incorporates so many data sources,” said Andrey Rzhetsky, senior author.

The new detailed surveillance could be a kind of “weather forecast for the flu,” they said. Tactical roadblocks against the epidemic where possible, before it spreads too far between states.

“For example, if flu-like symptoms are being report in one county, you could tell people in neighboring counties to stay away from crowds, or you could focus vaccination efforts in certain places in advance,” added Rzhetsky.