Data related to the PNAS paper. Accessed on Nov 14, 2015.
load_data(type ="extdata", ili.weighted =TRUE)
Arguments
type: the type of the data to be loaded. If type=="extdata" it loads the data to reproduce the PNAS paper, and if type=="athdata" it loads the data to reproduce the CID(?) paper.
ili.weighted: logical indicator to specify whether to load weighted ILI or not, if FALSE unweighted ILI is loaded.
Returns
A list of following named xts objects if type=="extdata"
GC10 Google Correlate trained with ILI available as of 2010. Google Correlate has been deprecated by Google as of Dec 2019 and is no longer publicly available.
GC09 Google Correlate trained with ILI available as of 2009.
GT Google Trends data for search queries identified using Google Correlate. Not directly available online, you have to manually input query terms at https://trends.google.com/trends/
A list of following named xts objects if type=="athdata"
GT Google Trends data for search queries identified. Not directly available online, you have to manually input query terms at https://trends.google.com/trends/
ili_idx the indexing information that includes the week number and year number, the date of ending Saturday, and the season number Available online at https://www.cdc.gov/flu/weekly/
ATH Athenahealth data that includes the proportion of "Flu Visit", "ILI Visit", and "Unspecified Viral or ILI Visit" compared to total number of visit to the Athenahealth partner healthcare providers.
ili_unrevised Historical unrevised ILI activity level. The unrevised ILI published on week ZZ of season XXXX-YYYY is available at www.cdc.gov/flu/weekly/weeklyarchivesXXXX-YYYY/data/senAllregtZZ.html
Parse and load CDC's ILI data, Google Flu Trend data, Google Correlate data trained with ILI as of 2010, Google Correlate data trained with ILI as of 2009, Google Trend data with search terms identified from Google Correlate (2010 version).
Each week ends on the Saturday indicated in the xts object
Google Correlate data is standardized by Google, and we rescale it to 0 -- 100 during parsing. Google Trends data is in the scale of 0 -- 100.
Yang, S., Santillana, M., & Kou, S. C. (2015). Accurate estimation of influenza epidemics using Google search data via ARGO. Proceedings of the National Academy of Sciences. doi:10.1073/pnas.1515373112.