On the night of 6th of April, 2020, I started analyzing COVID-19’s data from European Centre For Disease Prevention and Control (ECDC) to understand the abrupt rise in the number of reported cases within the United States and get a bigger picture of what’s going on. After looking at the data and analyzing it to the best of my abilities, I made some prediction based on the data. On 7th of April, I published my analysis and predictions in a blog, COVID-19 Part 2 – Situation In The United States.
In this blog, I’m going to analyze what I got right and what I didn’t. But, before I begin, let’s take a look at what I predicted.
Predictions
At the time of publishing that blog, the United States had a total of 337,635 cases reported. Based on the data available to that day, I made two predictions.
- The United States will report its daily maximum cases on 12th of April, 2020. This is infamously known as the peak.
This peak or more appropriately the peak of the Bell’s curve was all over the news. When experts on any news segment insisted the public to follow shelter-in-place or lockdown rules to flatten the curve, they meant that we need to keep the amplitude (the maximum height) of the curve as low as possible, thereby making it stretched and flat.
- My second prediction was that the United States will register 1 million cases by the end of April.
The output of the second Python script on my previous blog suggested that the 1 million mark will be reached on Day 92 or 22nd of April. I calculated the error for my calculated model, corrected it and projected the correction on the prediction to find the new date to be 28th of April. I rounded it off and published the blog with an ETA of the end of April, instead of the original 22nd April or the calculated, 28th April.
Now that April is over, lets take a look at how much of it was correct.
The Peak
Well, as the data is evident, the peak was reported on day 81 after the first reported case, i.e 11th of April as opposed to my predicted date, the 12th of April.
If you noticed that there’s a higher spike in the graph after day 81, you’re absolutely correct. But, in my defense, it was an outlier and wasn’t part of the Bell’s curve. Here, in the graph below, you can see that it was registered on day 92 or the 22nd of April.
Touching the Million Mark
Speaking of my second prediction, The United States touched the 1 million mark on 29th of April. Neither on the carefully calculated 28th nor on the carelessly rounded off 30th. So, I missed it by a day too. And, a day on both the sides! Looks like nature has its own way of embarrassing me! 
Conclusion
As discussed in my earlier blog, I calculated these prediction based on the gradual decline in the percent delta and here’s how it all went. If you’ve read the blog where I published my predictions, you’d be familiar with this graph.
Overall, my estimates were close. Because it’s what happens in the real world that gets reflected in the data and not the other way around, it’s impossible to predict these things accurately.
At the end, I can only say that my predictions were,
“A day too late and a spike too short!”