On The Grind

[datahurricanelogs]

Visualizing final results

Starting with the importing of the necessary libraries, choosing a theme and making sure our text gets printed out nicely on the plots:

Afterwards, making a histogram on the left and a boxplot on the right, showing median, IQR, and outliers (one plot per variable):

For a finish, using matplotlib to create a scatterplot/regression, including a Pearson correlation.

The results are as follows:

The results from this analysis do not find any significant correlation between employment rate and alcohol consumption.

Everyone drinks!

Thank you for being on this journey with me~

[datahurricanelogs]

Even more frequency distributions

This would serve as a continuation of my last post concerning the frequency distributions.

i continue my code by cleaning the columns and making sure I only leave in the values that are ≥ 0 for alcohol consumption and [0, 100] for employment rate.

Then I add two boolean flags: True if the cleaned value is not missing, False otherwise.

Then print everything:

Afterwards, I convert continuous variables into easy, human-readable groups (into quartiles) with the help of bins:

Afterwards, I define a helper that prints a clean frequency table for categorical data (or anything treated as categories plus a row for NaN because dropna=False):

Finally, just print everything:

The results are as follows:

This serves as a very nice continuation of everything we have done thus far.

See you in the next post~

[datahurricanelogs]

Frequency Distribution of Gapminder dataset

As per my last post, I did a frequency distribution between employrate and alcconsumption using Python.

This code: loads the file, makes sure both variables are numeric, builds a binned frequency table (counts, %, cumulative %) and prints missing values.

I have also created a version full with comments explaining how it works:

The results we get is two clean frequency distribution tables:

This type of analysis is a small but crucial step we take before diving deeper into our analysis.

See you in the next post~

[datahurricanelogs]

From the Office to the Pub: Alcohol Consumption and Employment Rate Linkage

After looking through the codebook for the Gapminder study, I noticed myself particularly interested in the connection between alcohol consumption and employment rate. So I decided to use exactly those corresponding variables (alcconsumption, employrate) in my own codebook.

My interest stems from both alcohol dependent people dealing with their unfortunate lack of employment, as well as those dealing with their unfortunate poor working environments. There are a lot of studies that have already tackled the effect of unemployment on alcohol consumption. It has been found to have positive links in alcohol consumption rate (Popovici et al., 2013). De Sio et al. (2020) have also found that office work reduces the chances of a person developing an alcohol dependence.

What I want to take a look at is how alcohol consumption and employment rates differ across countries and how it correlates with low and high employment rate regions.

Popovici, I., & French, M. T. (2013). Does unemployment lead to greater alcohol consumption?. Industrial Relations: A Journal of Economy and Society, 52(2), 444-466.

De Sio, S., Tittarelli, R., Di Martino, G., Buomprisco, G., Perri, R., Bruno, G., ... & Cedrone, F. (2020). Alcohol consumption and employment: A cross-sectional study of office workers and unemployed people. PeerJ, 8, e8774.