twiecki / covid19 Goto Github PK

View Code? Open in Web Editor NEW

156.0 15.0 57.0 27.91 MB

Analyses about the COVID-19 virus

License: BSD 3-Clause "New" or "Revised" License

Jupyter Notebook 99.98% Dockerfile 0.01% Shell 0.01% Python 0.02%

covid19's Introduction

covid19

Analyses about the COVID-19 virus.

Running the covid19 project in local Docker

To run this project on your local docker install, simply execute the ./run.sh command in Terminal.

In the terminal session a link for the Jupyter Notebook Server will appear, alongside your token:

    To access the notebook, open this file in a browser:
        file:///root/.local/share/jupyter/runtime/nbserver-8-open.html
    Or copy and paste one of these URLs:
        http://103cc22c52c6:8888/?token=<your_very_long_unique_token>
     or http://127.0.0.1:8888/?token=<your_very_long_unique_token>

Copy and past the link in your browser to access the Server.

Running the covid19 project notebooks on mybinder

If you don't have access to a local Docker or Jupyter Server, you can run the notebooks in an interactive session on mybinder:

https://mybinder.org/v2/gh/twiecki/covid19/master

covid19's People

Contributors

Stargazers

Watchers

Forkers

mikiobraun burkesquires existeundelta huyng abdelaty bensonlee5 yubozhao elianomarques nmlemus mitchjablonski snowdj raulrc fabiononato rajkrishnamurthy svenvanpoucke tlmnb zaxtax joemccann peter88tom avilher aseyboldt rvkempf simoneperazzoli nograpes dagalindo strayswonderland emitten89 jzajpt brettmontague alex92rus jingyuzhao-math mpeirlinck ephes karlthiele xiangze vtuyishime ajrader donnellymjd jeromeku kant yhogiwahyudi rebeltat pkjmesra ruinedsubmartingale tweetz0r mishrasushruti99 fagan2888 noellebr ronald-feng cactus1549 olgadk7 roesta07 hercules261188 gallowayj yishingene raamraam jamesmbaazam

covid19's Issues

Some explaination

Hello,

I am missing some explaination for the bayes part (covid19_growth_bayes.ipynb)
What are the underlying assumptions of the prediction?

Regards,
Hendrik

df_sign

1-this data frame has countries less than the whole list, based on what you filtered the countries?
2- why did you eliminate China?

Altair chart not showing

Yesterday, I can run all of the cells.
But, today I can't show the altair chart. Please, give me some advice.
Thanks

covid19_sir.ipynb errors on first cell

I haven't seen the dependency sunode before. It is not pip installable. I tried pip installing by pointing to the GitHub repo, and by cloning it:

git clone https://github.com/aseyboldt/sunode.git
pip install ./sunode

But they both error out. Could you explain how to install sunode? Thank you.

Bad arguments in plot

Love this primer. It's really insightful.
post_pred is a dict with obs key with this structure: {'obs': array([[5.71667615e+02, 1.34905178e+03, 5.25019985e+00, ...]])}

This won't render in the plot: post_pred[country].T

fig, ax = plt.subplots(figsize=(10, 8))
ax.plot(post_pred[country].T, color='0.5', alpha=.05);
ax.plot(confirmed, color='r')
ax.set(xlabel='Days since 100 cases', ylabel='Confirmed cases', title=country);

Why drop the cruise ship data?

See title. China is dropped but the cruise ship is also dropped without explanation.

Simplification proposal

Hi @twiecki,

I don't think there's a need to fully clone the John Hopkins repo. It is in fact sufficient to pull the raw data as follows:

import requests

...

base_url = 'https://raw.githubusercontent.com/CSSEGISandData/COVID-19/master/csse_covid_19_data/csse_covid_19_time_series'

...

def load_timeseries(name):
    url = f'{base_url}/time_series_19-covid-{name}.csv'
    csv = requests.get(url, proxies=proxies).text
    df = pd.read_csv(io.StringIO(csv), index_col=['Country/Region', 'Province/State', 'Lat', 'Long'], parse_dates=True)
    ...

Note that I have split the reading of the CSV text and the pandas parsing into two steps because I'm behind a corporate firewall and need proxies. You could combine the two into one by directly specifying the url in the call to pd.read_csv().