How Does Outbreak Prediction Work? | Tutorial with COVID-19 Data

This month's AI 101 focuses on outbreak prediction using COVID-19 data! We're visualizing exactly why flattening the curve is so important, and how social distancing accomplishes that.

References:

Python Tutorial Used to Create Colab Example:

[trainsonplanes]

I didn't even notice that you changed your channel name until you mentioned it. Well, it is easier to pronounce now, so there's that. Love your vids!

[nerdlarge]

Nice demonstration. Maybe once the pandemic matured and the data becomes more reliable, you can compare COVID19 to 2009 H1N1 strain to give people a better perspective on how different this pandemic is to the previous one.

[zhubajie]

Just to keep my mind busy while social distancing, I have been plotting the Florida confirmed case curve on a logistics curve (adjusting the maximum to give the best rsq of the equation rearranged to be linear) just with a spreadsheet. The maximum is of course increasing as one gets more data in as we're still in the exponential phase. It will be interesting how accurate the logistic curve model will be once the growth rate slows. From even lower number the plateau value yesterdays was 1411, today 1862, probably higher tomorrow.

The problem to is people are Spring Breaking here so its not a closed system but the beaches and/or beach parking is closing so people are getting fewer on the beach transferring their viruses then spreading them throughout the country. Wish the governor had put a stop to being on the beach immediately (though he did close the bars but they get around that my serving more food to qualify as a restaurant). He left it up to the individual localities to close beaches or not. Dumb.

[andra]

i've just come across this video now, but this was an amazing explanation, ty!!

[mahaletealemu]

This explained a lot . Thank you. You are amazing!

[andrewpynch]

Love your videos. Glad I found this channel

[cyncynshop]

Great work!
Coronavirus is recoverable, but having weakened lungs is *permanent.* Recovering from Corona does not mean it wouldn't leave lasting consequence on your body.

I am planning to make a graph to compare this graph with the flu cases from last year December.
Last winter the number of flu cases increased rapidly along the number of pneumonia deaths, so I do want to compare the Winter 2019 to Spring 2020.

[ozzyfromspace]

Sadly, the model couldn’t account for politics. We just hit 5 million confirmed (not actual, which is a lot more). Nice video, Jordan! I’m an instant fan ofc 😊🙌🏽🎊 #Subbed. It’s awesome to see intelligent women doing their thing ☺️

[shantanukris]

Thats a good explanation. Have you looked at fitting a logistic function - especially if the growth in new cases has reduced or is close to 1?

[brown]

"Many of you have been working from home for a couple days"

Ahh, that aged...

[gecarter]

Great video. Anybody know how to resolve this: Running read-only notebooks is not supported. Make a copy of this notebook or switch to playground mode.

[ibrahimimano]

Excellent explanation. Congratulation . I have a question what if we integrate the real data and use it to predict the days when the covid-19 will end.

[vnceigz]

• You've probably already heard by now, but IBM's Summit supercomputer at the DoE has identified 77 chemicals that could help stop the spread of SARS-CoV-2. I imagine it probably used at least some A.I.
• 10:57 - I've also seen young people complaining that their parents aren't taking it seriously. Everyone's blaming everyone. :-|
• 11:23 - Maybe it's the pessimist in me, but I imagine this could be used as an attack by some misanthropes…