NA values check in Challenge Library

[bestill3]

I checked the NA values against the challenge json files defined in library. The steps are:

1. The hubs checked are flusight-forecast-hub and rsv-forecast-hub. The columns used in their time series data are target_end_date, location, and observation.
2. Given the challenge json file uses reference_dates, their target_end_dates are obtained by the formula target_end_date = reference_date + horizon * 7 days.
3. Filter the ground truth of time series data by using combination values of target_end_dates and locations that are calculated/contained in the json file.
4. Count the NA values in the filtered ground truth.

Here are the findings:

Challenge Name	Hub Name	Number of NA values in Column Observation
epb_flu_inchosp_2023-2024_v1.json	flusight-forecast-hub	258
epb_flu_inchosp_2024-2025_dev.json	flusight-forecast-hub	0
epb_flu_inchosp_2025-2026_dev.json	flusight-forecast-hub	0
epb_rsv_inchosp_2025-2026_dev.json	rsv-forecast-hub	0

An example of the NA in flusight-forecast-hub time series data is: target_end_date=2024-05-18; location=25. It could link to epb_flu_inchosp_2023-2024_v1.json reference date=2024-04-27; horizon=3. A screenshot is also provided below.

@jcblemai @emprzy , it seems that the epb_flu_inchosp_2023-2024_v1.json has a fair amount of NAs, what would be the plan for this challenge?

[emprzy]

@bestill3 can you please paste in a summary statement (locations, reference dates, horizons) for the NA observed values in epb_flu_inchosp_2023-2024_v1 so i can get an idea? instead of a single example

[bestill3]

Please see the attached summary.

Challenge_NAs_Summary.xlsx