Homework 02

Youth Tobacco Survey: results

This week, you will continue with the Youth Tobacco Survey (data available here).

Here is a compiled dataset from all provided years: yts_summarized.csv, that contains the following variables for each combination of year, sex, and grade:

• n: sample size (i.e., number of students in this combination of age, sex, and year)
• p_cig: proportion of students who report having ever tried a cigarette
• p_vape: proportion of students who report having ever tried an e-cigarette
• first_age: mean reported age at which students first smoked, of those who have reported smoking
• num_days: mean number days on which smoked cigarettes of the past 30 days, of those who have reported smoking
• p_will_smoke: proportion of students who say they will smoke a cigarette in the next year
• p_harmful: proportion of students who say that they think the smoke from other people’s cigarettes is harmful to them

The compiled dataset was produced by this script: yts_summarize_data.py

Assignment: You should upload to Canvas a self-contained ipynb document, that should be readable as a report: in other words, by selecting View > Collapse All Code, I should have something that looks and reads like a report (ignoring the code).

Please make sure to address the following points. Note that some of these can be re-used (at least in part) from HW1.

• an introduction describing the data: how it was collected, what it was collected for, and where it was downloaded from
• a precise description of each of the variables above, and how it was extracted from the dataset
• a description, including plots, of how each of the variables (other than n) differ across years, grades, and sexes
• a conclusion giving the big-picture takeaways

In particular, please include the following elements:

• at least one plot with ‘grade’ on the x-axis, a smoking-related variable on the y-axis, and with separate lines per year
• at least one plot with values aggregated across grade and/or sex
• at least one plot with separate facets by reported sex
• at least one plot showing the difference in the smoking-related variable between reported sexes
• identification of at least one error in the summarized dataset (this should be visible as a nonsensical pattern in the plots) and explanation of where the parsing code (summarize_data.py) went wrong

Furthermore, please:

• omit points in the plots with small sample sizes (i.e., remove the noisy/unreliable categories).
• label all axes, etcetera