Economists all over will tell you that all recessions are different. Some economists believe that there will be one in the second half of 2023. Our AI cannot tell you that, yet. What we can tell you, however, are certain indicators of past recessions. Elevated rates of unemployment is one such indicator that is well documented. However, searching and presenting such data points is not necessarily easy.
With Vanna, we are able to query data using plain English that our AI will then convert into SQL. Plain and simple.
For starters, we obtained data from the Bureau of Labor Statistics via the Snowflake Marketplace. The Bureau of Labor Statistics publishes data related to labor markets in the US.
We then asked Vanna how the labor markets looked between 2007 and 2022. After all, there was the Great Recession between 2008 and 2011 and then Covid-19 between 2020, to the present-day. Instead of asking Vanna what the unemployment rates were, we want to smooth it out into 3-month running averages to account for seasonal spikes.
According to Vanna, this is what they looked like:
Now that we have a summary of the unemployment rates stretching back from before Justin Bieber became a thing, we also want to ask whether the two recessions were roughly similar in terms of levels of unemployment.
We asked Vanna to highlight the states with the 10 biggest differences in mean unemployment rates over a 3 year period from 2008 to 2010, and 2020 to 2022.
And voila! We got an SQL query that would have taken an analyst some time to write, and a pretty informative chart that seems to indicate that there is a significant difference in unemployment rates at the state level.
In a throwback to Stats 101 in college, we might want to do a paired t-test of the mean unemployment rates to see if there is a difference between the two time periods, and how statistically significant it is. We can of course download the data and do it in Python. But why do it manually when we can ask Vanna to do it for us in a matter of seconds?
With a t-statistic of -11.20, and 52–1 = 51 degrees of freedom, we can plug these numbers into an online calculator, obtain the results, and reject the null hypothesis that the difference in mean unemployment rates between 2008 and 2010 and 2020–2022 was zero. That is to say, the mean unemployment rates among states were lower in 2020–2022 than in 2008–2010.
Even though Vanna AI might not be able to predict the future at the moment, it can distill complicated questions into queries without much effort on the part of the analyst. In a nutshell, Vanna gives the analyst more time to analyze, and less time spent writing repetitive queries. Connect with Vanna AI today to find out more!
