refactor jhu data processing to calculate inc signal

[elray1]

move processing from load_jhu_data function to the assemble-historical-jhu.R script

Create a new R function called calc_jhu_inc that has the code in these lines:

covidData/R/jhu_data.R

Lines 97 to 172 in feb5648

	tidyr::pivot_longer(
	matches("^\\d{1,2}\\/\\d{1,2}\\/\\d{2,4}$"),
	names_to = "date",
	values_to = "cum"
	) %>%
	dplyr::mutate(
	date = as.character(lubridate::mdy(date))
	)
	# summarized results for county level
	results <- NULL
	if ("county" %in% spatial_resolution) {
	county_results <- jhu_data %>%
	dplyr::filter(FIPS > 100) %>%
	dplyr::mutate(
	location = sprintf("%05d", FIPS)
	) %>%
	dplyr::filter(location < "80001") %>%
	dplyr::group_by(location) %>%
	dplyr::mutate(inc = diff(c(0, cum))) %>%
	dplyr::select(location, date, cum, inc) %>%
	dplyr::ungroup()
	results <- dplyr::bind_rows(results, county_results)
	}
	# summarized results for state level
	if ("state" %in% spatial_resolution) {
	states_to_keep <- c(
	"Alabama", "Alaska", "American Samoa", "Arizona", "Arkansas",
	"California", "Colorado", "Connecticut", "Delaware",
	"District of Columbia", "Florida", "Georgia", "Guam", "Hawaii", "Idaho",
	"Illinois", "Indiana", "Iowa", "Kansas", "Kentucky", "Louisiana", "Maine",
	"Maryland", "Massachusetts", "Michigan", "Minnesota",
	"Mississippi", "Missouri", "Montana", "Nebraska", "Nevada",
	"New Hampshire", "New Jersey", "New Mexico", "New York",
	"North Carolina", "North Dakota", "Northern Mariana Islands",
	"Ohio", "Oklahoma", "Oregon", "Pennsylvania", "Puerto Rico",
	"Rhode Island", "South Carolina", "South Dakota", "Tennessee",
	"Texas", "Utah", "Vermont", "Virgin Islands", "Virginia",
	"Washington", "West Virginia", "Wisconsin", "Wyoming"
	)
	state_results <- jhu_data %>%
	dplyr::filter(Province_State %in% states_to_keep) %>%
	dplyr::mutate(location_name = Province_State) %>%
	dplyr::group_by(location_name, date) %>%
	dplyr::summarize(cum = sum(cum)) %>%
	dplyr::group_by(location_name) %>%
	dplyr::mutate(inc = diff(c(0, cum))) %>%
	dplyr::ungroup() %>%
	dplyr::left_join(
	covidData::fips_codes %>% dplyr::filter(nchar(location) == 2),
	by = "location_name"
	) %>%
	dplyr::select(location, date, cum, inc)
	results <- dplyr::bind_rows(results, state_results)
	}
	# summarized results for national level
	if ("national" %in% spatial_resolution) {
	# because we don't filter on states_to_keep as above, we are off by a total
	# of 3 deaths attributed to Diamond Princess.
	national_results <- jhu_data %>%
	dplyr::group_by(date) %>%
	dplyr::summarize(cum = sum(cum)) %>%
	dplyr::ungroup() %>%
	dplyr::mutate(
	inc = diff(c(0, cum)),
	location = "US"
	) %>%
	dplyr::select(location, date, cum, inc)
	results <- dplyr::bind_rows(results, national_results)
	}

. We can save only the calculated incidence, not the cumulative values (as they can be reconstructed from the incidence). So the new function should return a tibble or data frame with columns location, date, inc.

In assemble-historical-jhu.R, at

covidData/code/data-processing/assemble-historical-jhu.R

Line 18 in feb5648

function(filename) suppressMessages(readr::read_csv(filename)))

and

covidData/code/data-processing/assemble-historical-jhu.R

Line 32 in feb5648

function(filename) suppressMessages(readr::read_csv(filename)))

, call the new function to process those inputs. After this point, the jhu_deaths_data and jhu_cases_data structures should be tibbles with columns issue_date and data, where in a given row, data is the tibble or data frame created above with columns location, date, and inc corresponding to the data for that issue date.