Technical requirements for data for forecasting

1. On setup: The location of the unit(s) for which a forecast has to be produced. If the forecast is for multiple units then a list of locations. Preferably, a GPS coordinate for each unit, but an address can be applied.

2. Data on the number of hourly contacts as far back in time as possible (with a minimum of three years). Data should be provided as a CSV file (separated by “,” and “.” as decimal point) in the UTC time (ISO 8601-1:2019, see https://en.wikipedia.org/wiki/ISO_8601), e.g., YYYY-MM-DDThh:mm:ssZ where:

1. YYYY = year (2024)
2. MM = month (05)
3. DD = day (10)
4. hh = hour (11)
5. mm = minute (36)
6. ss = seconds (15)
7. T = seperates date and time
8. Z = indicates Zulu time = UTC.

An example could be:
UtcTime , Value
2023-08-11T06:00:00Z , 0
2023-08-11T07:00:00Z , 3

For the forecast to be given continually, it is required that these data are updated frequently (preferably as soon as data becomes available, e.g., hourly).

3. Information on events that could affect the forecasts:

1. Event period: Start (UTC time) and end (UTC time) as specified by ISO 8601-1:2019.
2. Name of the event.
3. Locality: How local is the event? Country, NUTS-1, NUTS-2, NUTS-3, LAU, or Specific, where NUTS levels and LAU are defined by Eurostat (https://ec.europa.eu/eurostat/).
4. Location: Where the event takes place: as precise a location as possible according to the locality given:

   a. Country: ISO 3166-1, Alpha-2 (see https://en.wikipedia.org/wiki/ISO_3166-1). For Spain it is “ES”.

   b. NUTS-1,…, NUTS-3, Nomenclature of Territorial Units for Statistics (NUTS, see https://ec.europa.eu/eurostat/web/nuts), for instance “Comunidad de Madrid” is at NUTS-2 level and has the code ES30 while “Madrid” is at NUTS-3 level and has code ES300.

   c. LAU, Local Administrative Unit (LAU, see https://ec.europa.eu/eurostat/web/nuts/local-administrative-units), where the LAU code for (as an example) Aranjuez is 28013 (and it is a part of Madrid which has NUTS-3 code ES300)

   d. Specific, the GPS coordinate.

   If the event occurs on multiple locations, then a list of locations can be given.

5. Type of event: This is a set of tags describing the event. For instance an event type could be the two tags “soccer” and “El Clásico”, or another could be “festival” and “Bonfires of San Juan”.
6. Previous occurrences (e.g., previous years)
7. Later occurrences (e.g., the next years)

4. Information on organizational changesthat could affect the forecasts. E.g., change in population size, change in procedures etc.

Our system has the following technical requirements for performing the forecasts:

1. Access to the internet
2. Computational power and storage similar to a large desktop computer.
3. An OCI-compliant (https://opencontainers.org/) container based environment such as Docker (https://www.docker.com/) or Podman (https://podman.io/).

Our system has the following technical requirements for developing the forecasts:

1. Access to the internet
2. Access to computational power similar to 8 cores, 64 GB memory and 1 TB storage per concurrent model training.
3. An OCI-compliant (https://opencontainers.org/) container based environment such as Docker (https://www.docker.com/) or Podman (https://podman.io/).