Hi, just reminding you that I'm a teacher who nerds out about data and therefore made assignments about COVID for my students. I like these assignments enough to share them and remember them so here you go.
Welcome to day two (of three) regarding COVID data. Today I have some background info for you and some questions (yours and mine) that I'd like you to try and answer.
I shared Scott Bond's Utah graph from April 28 last time. I'd like you to understand a bit about his data collection and graph before we continue, so I'll use his graph of the COVID numbers from Italy to describe.
Scott always graphs three lines. The first line is the raw data, AKA the actual number of new cases each day. For the Italy (and Utah) graph, that is represented by the grey and most jagged line.
Now, Scott noticed that the numbers from day to day sometimes vary greatly and for unrelated reasons to COVID spread (for example, Utah's testing always slows on Saturday and Sunday, which means the Monday numbers tend to be higher). So he averages the last five days of data and plots those points as a new line. This smooths out the data a bit, and on the Italy graph you can see it as the light green line (light purple on Utah's graph). This line is still bumpy, but it gives a better snapshot of whether case numbers are increasing or decreasing over time.
Finally, based on that 5-day average line, Scott writes an equation for a bell-shaped line that will best fit the data. He just keeps guessing and checking until he finds one that fits nicely, and then as the data changes over time, he adjusts the final graph. Check out this graph from April 20 where Scott realized the right side of the curve should be less steep than the left side of the curve for Italy:
Do you see how the jagged lines started to be too far away from Scott's smooth green curve? When he could tell this was happening (it takes a few days of data to know there is a deviating pattern), he wrote a new equation (the red line) that would better fit. The red line in this graph is the green line you saw in the first graph...and the data since April 20 has stayed much closer to this new smooth line!
Okay, so now let's go back to the Utah graph which I shared with you last time:
Your assignment today is to practice reading this graph. Respond to these questions and recognize that there may not be a particular right answer to each of them. Use the graph to help you!
1) Overall, how well do you think Scott's purple line fits the COVID numbers for Utah? Explain.
2) If you were Scott, would you review and/or rewrite your equation for the smooth purple line? Why or why not?
3) Describe the story this graph tells about COVID-19 in Utah. (Use the raw data and/or the 5-day average if you disagreed earlier with Scott's smooth purple line.)
4) Based on the graph, what do you predict happened with Utah's data between April 28 (the end of this graph) and May 7 (today)? (You are welcome to look up this data on Scott's FB profile if it interests you. We will also see that data next time.) What do you predict will happen after today?
5) What are the advantages and disadvantages of having a mathematical model that predicts future COVID numbers? Name and explain a couple.
Optional 6) What new questions do you have about the graph?
No comments:
Post a Comment