Get historical reproduction numbers for Lithuania

[ramashka328]

Hello developers,

thank you for a great analysis! I would like to get an effective reproduction numbers for Lithuania starting from 2020-02-17. Is it possible to get this kind of information using your provided resources?

Thanks in advance!

[joeHickson]

Hi @ramashka328 ,

sorry for the slow response. There are a few options:

1. Change the data window from 12 weeks (https://github.com/epiforecasts/covid-rt-estimates/blob/master/R/utils.R#L75) and run just Lithuania from the global datasets using -i "cases/Lithuania, deaths/Lithuania". You could do this properly using expose the 12 week window as a parameter #102 if you fancied tackling a ticket or just hack it in for your own use locally.

2. within the next few months we are thinking we will change infrastructure and able to run the full historic datasets every so often (this is wishlist so no guarantees). This might include the global cases / deaths datasets that would include a top level number for Lithuania.

3. find a data source that gives breakdown within Lithuania and incorporate it into the project formally - https://github.com/epiforecasts/covid-rt-estimates/blob/master/SMG.md should give some guidance but do ask for more advice if you need it. In combination with 1)/expose the 12 week window as a parameter #102 this would give you more detailed figures for the country but may be more than you need.

4. find some relevant data, massage it into the right format and mash it into the script here: https://github.com/epiforecasts/covid-rt-estimates/blob/master/R/update-regional.R#L121 - this is a quick and nasty version of 3 but absolutely fine for one off work. You are then really using the EpiNow2 library but you can use this project as an outline of how to call it (copy/paste for the win).

I hope this helps

Joe

[seabbs]

Hi @ramashka328 thanks for opening this,

If taking option 4 you might find the docs for EpiNow2 helfpul: https://epiforecasts.io/EpiNow2/dev/

@joeHickson (hope you had good AL) - we are getting increasingly freqeunt requests for historical Rt values so it might be a good idea to bump historical Rts up the food chain in terms of priority.

Sam

[sbfnk]

+1 for historical Rt trajectories as a priority.

[seabbs]

Whilst we work on historical estimates here is a case study from generating Rts from scratching for a country in the ECDC data: https://gist.github.com/seabbs/163d0f195892cde685c70473e1f5e867

Note this needs to be adapted to use realistic delays for the country of interest.