Data¶
The number of people developing symptoms daily are compared to data obtained from several sources at the national, state, or regional levels [JHUCOVID19] [NYTCOVID19]. We found that, for some states or regions, the reported daily counts exhibited a significant amount of noise. This is caused by variation in testing capabilities and sometimes by how data is aggregated from region to region over the coarse of a week. To filter the noise observed in daily case count data, we use of 7-day rolling averages. Time series of daily counts, raw data with black symbols and filtered with red symbols, are presented in Fig. 3. We employ filtered data in the examples shown in this manual.