Data processing
Data pipeline
The data pipeline transforms raw data into data suitable to be used as input for the training of forecast models.
Data acquisition
Data curation
Data transformation
Data integration
Data reduction
Data preparation
Data analytics
A set of statistical techniques to extract information from data.
Input features for COVID-19
The most common feature to be forecasted is the time-series of deaths caused by COVID-19.
Input features |
|---|
Cases (infected) |
Deaths |
Recovered |
Mobility |
Temperature |
Humidity |
Air Quality Index (AQI) |
# of Vaccinated |
Population vaccinated % |
# of people tested |
Negative tests |
ICU occupancy % |
ICU occupancy by group age |
# of people on ventilation |
# of hospitalized |
Excess Deaths |
Population Density |
Death rate (deaths / population) |
Infected rate (cases / population) |
Mortality (deaths / infected) |
Reproduction number (R_0) (transmission rate) |
Local holidays |
Engine searches |
Sentiment data from social networks |
Device exposures (mobile data to measure social contact) |