Today we were tasked with looking at the American Time Use Survey (ATUS) which measures a range of activities that people spend time doing. Trying to interpret the data was quite difficult initially as the structure provided by the Bureau of Labor Statistics was not very tableau friendly. Luckily for us, we have an ace up our sleeve… Alteryx!
Most of it involved repeating the same logic and creating multiple joins. Essentially we had tables that could be matched together based on unique codes. Firstly we created the main table that contained all the data in a coded format which we could then use to reference other tables using the data dictionary on the site. Now we just had to create separate workflows for each column we wanted to decode and join it back based on the column code to the main table to get our corresponding values. Once the joins were complete the workflow was more or less done just a little rearranging and cleaning was required using the cross-tab and select tool to structure the data. Now that we had 1.27 million records outputted it was time to explore the data in Tableau! In hindsight, I think the process would have been much more efficient if I decided on what factors I wanted to focus on before pulling in every piece of data I could. Thanks to Luisa, Soha, and Laine for their help with the joins!
Skimming through the list of activities in the survey, there was one that I particularly wanted to focus on which was time spent on job hunting and I wanted to see whether this correlated with the unemployment rate, which I supplemented the data for. However, being survey data I quickly realised that there would be data quality issues and that any insights coming from the data would likely be skewed to some extent. Nonetheless, it was interesting to look at the gender and race breakdown of time spent job hunting. Surprisingly, the older generation seemed to be more active job hunters. Check out the viz to find out more!