Covid-19 vaccine forecasting story

[jack89roberts]

Summary

Forecasting when the UK will fully vaccinate the adult population with data collected from the API of the government's Covid-19 dashboard, and using fairly simple approaches such as "when will the population be vaccinated if doses continue to be given at the same rate as the last week?"

Fixes #129

List of changes proposed in this PR (pull-request)

• New story notebook (stories/covid-vax/covid_vax.ipynb)
• environment.yml conda environment for dependencies (which are minimal, only requests, pandas, numpy, matplotlib and seaborn) - not sure how dependencies are handled for the stories?
• Backup data file 20210504_data.json in case the API changes/is unavailable in the future.

What should a reviewer concentrate their feedback on?

• [x] Length and level of detail ok? I feel like it got quite long for something that's actually pretty simple. E.g. is the amount of time describing the code ok - I was trying to make it as accessible as possible but might have gone over the top in places.
• [x] Approach reasonable? In particular the forecast with the uncertainty band? It's maybe not the most robust approach but I was trying to avoid adding the complexity of using an "actual" forecasting algorithm.
• [x] Is the explanation of how the forecasts decide whether doses will be given as first or second doses clear? And of calculating the delay between 1st and 2nd doses?
• [x] Does the code make sense?
• [x] Everything looks ok?
• [x] Cut off date for the data is currently set to 4th May. This could be updated closer to the time of publishing, but it's also true that the forecasts become less interesting as time goes on and the vaccine programme gets closer to completion.

Acknowledging contributors

• [x] All contributors to this pull request are already named in the table of contributors in the README file.
• [x] The following people should be added to the table of contributors in the README file:

[review-notebook-app[bot]]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[crangelsmith]

@all-contributors please add @jack89roberts for blog

[allcontributors[bot]]

@crangelsmith

I've put up a pull request to add @jack89roberts! :tada:

[crangelsmith]

@all-contributors please add @jack89roberts for idea

[allcontributors[bot]]

@crangelsmith

I've updated the pull request to add @jack89roberts! :tada:

[crangelsmith]

@all-contributors please add @billfinnegan for review

[allcontributors[bot]]

@crangelsmith

I've put up a pull request to add @billfinnegan! :tada:

[jack89roberts]

@crangelsmith @billfinnegan Thanks again for all your comments! I believe I've now implemented them all, but will give it another read-through tomorrow to double-check. This includes adding markers to show the latest actual values from today in the last plot @crangelsmith , which could be updated again on the date it will be published if you like. I've replied to some of the other suggestions separately above.

I'm happy to help with any of the next publishing steps too - we might find it needs a bit more tweaking to fit the "two narratives" idea, for example.

[jack89roberts]

Hi again @crangelsmith @billfinnegan - I've had a final read through and made some last edits. Let me know if you have any comments on the changes, but otherwise I think it's ready 😄

[crangelsmith]

This is fantastic @jack89roberts , i'll try to give it a second read today. Do you think you could join our meeting on Wednesday? We could work together with @samvanstroud in getting it ready for publishing (e.g different colours for the two narratives, and making sure everything looks good in html, etc).

[jack89roberts]

Sure! There is a meeting about the RDS course 13:00-13:30, but I'm available Weds afternoon otherwise.

[samvanstroud]

Thanks all for writing and reviewing this story @jack89roberts @billfinnegan @crangelsmith ! Merging this now.