Consider creating covidcast meta tables

[sgratzl]

covidcast_source_info

• id ... PK ... same as source
• name
• description
• license? ... under which license are the signals of this source published?
• link ... url to docu

covidcast_signal_info

• id ...PK ... signal name, same as signal
• source ... PK ... source name, FK -> covidcast_source_info
• name
• short_description
• description
• is_smoothed: bool ... whether this signal is smoothed
• is_cumulative: bool ... whether this signal is a cumulative signal
• is_weighted: bool .... whether this signal is weighted -> see fb-survey weighted versions
• format: per100k, percent, fraction, count, other ... what does the value mean
• has_stderr: bool
• has_sample_size: bool
• category: early, public, late ... classification for hierarchy
• high_values_are: bad, good, neutral ... classification of meaning
• base_name: str ... what is the base_name of this signal, e.g., confirmed_incidence_prop, confirmed_cumulative_num will all have the same base_name of confirmed ... used to find related signals
• link ... url to docu
• time_label: str ... label which should be used for the time_values, e.g. Reported Date or Date
• value_label: str ... label which should be used for the values, e.g., Percentage, People, per 100 people, per 100,000 people, Fraction of population

[sgratzl]

closed with /covidcast/meta endpoint