1 of 9

COVID-Inspired Data Science Education through Epidemiology (CIDSEE)

Presenter: Jacob Sagrans

CSTA New England regional conference

November 12, 2022

Link to slides: https://bit.ly/CSTANE22CIDSEE

2 of 9

Agenda

Welcome/introductions (10 min)
Project overview (10 min)
Exploring 1918 flu data in CODAP (10 min)
CODAP/NetLogo simulation of spread of COVID under different conditions (10 min)
Closing/Q&A (10 min)

2

3 of 9

CIDSEE project overview

3

4 of 9

Data Detectives Clubs program components

Book and podcasts
CODAP/NetLogo activities (codap.concord.org)
Non-digital activities

4

At the heart of our project is our out-of-school program for middle school youth, which we call “Data Detectives Clubs.” To date, Clubs have been implemented in around 50 afterschool and summer camp programs around the US, reaching over 650 youth. Clubs take up to 20 hours, usually spread over several weeks or months. Educators are guided by a detailed program leader guide we have prepared.

At the center of each Data Detective Club is a middle school adventure novel titled The Case of the COVID Crisis, written by co-PI Pendred Noyce, with accompanying podcasts featuring the characters. Each Club session typically revolves around a specific book chapter, often with digital data activities in the Common Online Data Analysis Platform, or CODAP. There also activities involving a simulation tool called NetLogo that can be embedded in CODAP. We will get to explore CODAP and NetLogo in more depth shortly. In addition, there are non digital activities, such as an infection spread simulation game involving cups of water/vinegar and pH strips and a data sonification activity where youth use kazoos to “play” daily new COVID case graphs for different states (going up in pitch as cases increase and down in pitch as they decrease).

5 of 9

“Trailer” for The Case of the COVID Crisis

5

6 of 9

Clips of a Data Detectives Club in action

6

7 of 9

Exploring 1918 flu dataset in CODAP

Activity link: https://bit.ly/CCPittsFlu1918
When do the number of new infections peak?
When do the number of deaths peak?
Why do deaths peak at a different time than the number of new infections?
How would you describe the shape of the graphs? Increasing? Decreasing? Some combination of the two?
Pretend it is November 30th, 1918, when the graphs/data stops. Looking forward, when do you think daily new infections will be less than 40?

7

10 minutes – end by 11:30

Now we will do the first of two hands-on activities to try out digital data tools that we use in Data Detectives Clubs.

In this first one, we will use CODAP to explore data on the 1918 Flu, with a focus on comparing daily cases and daily deaths. This is an activity that complements chapter 2 of the book, where the characters visit Pittsburgh in 1918 to learn about the flu pandemic then. This is actual data from the 1918 flu pandemic in Pittsburgh, taken from a published article and entered into CODAP. Since this is an earlier chapter in the book, this activity is on the simpler side and serves as an introduction to CODAP.. And I just want to point out the illustration from this chapter showing women of the era sewing masks–yes, people knew to wear masks and avoid crowds to limit the spread of disease even back then!

So everyone will need to open the link on the slide to access the data in CODAP. Feel free to pair up if you want. Has anyone here used CODAP before–if so raise your hand! CODAP newbies may want to pair up with those of you with some experience.

What we are all going to try to do is make a graph showing the change over time in daily new infections in Pittsburgh in the fall of 1918, and then add daily deaths to this graph so we can compare cases vs. deaths, trying to answer questions such as when cases and deaths peak, why the peaks may be at different times, what the general shape of the graphs are, and how we might use the data to predict into the “future” (or really, into December 1918, after the graphs/data stop).

The instructions for the activity are included in the CODAP file at the link on the slide (Jacob: open the link and show what this looks like, also start doing the activity after a few minutes).

What did you find? Let’s try to answer the questions on the slide.

This activity gets youth to start thinking about time-series data as well as possible relationships between two variables (cases and deaths), and also to think about making predictions for the future (which is very complicated in this case, given the variability in the data).

Note that we do not typically dwell on deaths in Data Detectives Clubs, especially not currently with COVID, given the sensitivity of discussing this with youth.

10 minutes – end by 11:30

Paste into textbox in CODAP:

Activity link:

https://bit.ly/CCPittsFlu1918

Questions:

- When do the number of new infections peak?

- When do the number of deaths peak?

- Why do deaths peak at a different time than the number of new infections?

- How would you describe the shape of the graphs? Increasing? Decreasing? Some combination of the two?

- Pretend it is November 30th, 1918, when the graphs/data stops. Looking forward, when do you think daily new infections will be less than 40?

8 of 9

CODAP/NetLogo to simulate COVID spread

CODAP/NetLogo virus simulation link:

http://bit.ly/CCCODAPNetLogo

Parameters to use:

300 peeps (people)

Survival rate 50 percent

Duration of illness 21 days

R0 = 5

Try out different %’s starting immune:

0%, 25%, 50%, 75%, 90%

Fill out one row with your name and ending populations (number still alive) in this spreadsheet: https://bit.ly/CIDSEEspreadsheetCSTANE

8

10 minutes (end by 11:40)

For our second hands-on activity, we will use CODAP with NetLogo embedded in it to simulate the spread of COVID under different conditions, considering what happens to infections when different percentages of people are vaccinated or otherwise immune (from having had COVID). How long does it take for the simulation to run its course under various parameters, and how many people get sick? How variable are the results you get each time you run the simulation with the same parameters?

This is an activity we do near the end of the Data Detectives Clubs, as it is more complicated. Youth actually see a simpler version of this virus spread simulation early on in the Clubs, but now we are adding in the effects of vaccination, and also the possible variability over different runs of a simulation with the same parameters.

First I will show you how the basics of how the simulation works. Then everyone will try it on their own (or in pairs) and fill out a shared spreadsheet with your results. We will use the parameters on the slide here, trying different percentages starting immune. A couple things to note – we are using a survival rate of only 50%, meaning 50% of infected people will die. This is much lower than what COVID actually is. We are also using an R naught (or R0) of 5. An R naught of 5 means that on average, an infected person will infect 5 other people. An R0 of 5 is comparable to the R0 of the coronavirus delta variant we were seeing last fall. The only thing we will change from simulation to simulation is the % of people starting immune. If the model says 50% of people are starting immune, that means half of the people are incapable of being infected, which is an oversimplification, as it assumes vaccination (or previous infection) gives perfect immunity to every person, which is not the case in real life.

Okay, I will now show what this actually looks like in CODAP/NetLogo.

Jacob, keep the simulation open with the parameters.

After a few minutes look at the results in the spreadsheet.

Talk about the variation in the results.

Talk about what percentage people think would be a good percentage to vaccinate.

Mention that we talk with youth about the tradeoffs of increasing vaccination beyond a certain point (could be a high cost for only a few more people surviving–but what about ethics? We may want to make sure as many people as possible are vaccinated to limit any deaths/cases, even if the number is small).

Paste the following in a textbox in the simulation

CODAP/NetLogo virus simulation link:

http://bit.ly/CCCODAPNetLogo

Use these parameters:

300 peeps (people)

Survival rate 50 percent

Duration of illness 21 days

R0 = 5

Try out different %s starting immune: 0%, 25%, 50%, 75%, 90%

Fill out rows with your name and ending populations (number still alive) in this spreadsheet

https://bit.ly/3yvz3uy

10 minutes, end by 11:40 am

9 of 9

Closing/Q&A

Link to one-page project overview: https://bit.ly/3rbYBJP
Link to request our curriculum: https://bit.ly/DDGuideRequest
Link to buy the book: https://bit.ly/37dRfg8
Link to the slides again: https://bit.ly/CSTANE22CIDSEE
Contact me!: jacob.tumblehome@gmail.com

9

10 min (end by 11:50 pm)

Okay, so I gave you an overview of the COVID-Inspired Data Science Education through Epidemiology project and our Data Detectives Clubs, and we also did some hands on activities taken from the Clubs.

Does anyone have any questions at this point?

Before ending, I want to highlight I have put a bunch of resources on this slide for you to access in the future, as well as the logos of our project partners. Feel free to use anything I showed you today with your students.

Also, I want to highlight the first link on this slide, which is a sign up link for a middle school teacher focus group we are planning to have. We are interested in potentially expanding our curriculum into classroom learning. We would like to learn what it would take to do this from teachers who are potentially interested in using the curriculum. You will receive a $50 stipend as well as a copy of the Case of the Covid Crisis book. Sign up by November 20th at the latest, and feel free to forward to other teachers you know you who may be interested. If you sign up right now I can give you the book right now! Also I have some handouts with more information about the focus group if you want to take any, as well as a handout highlighting NGSS, Common Core Math, and other standards that we touch on in the curriculum.

Thank you!