March 2022 Datathon project: Pub visitation in Chicago around St. Patrick's Day

Notebook: Google Colab


St. Patrick’s Day has been widely celebrated in the US. Besides the parades and dying the river, going to pubs is also a big part of the celebration. Despite its popularity, St. Patrick’s Day is actually not a national holiday, which means that most people still need to work on that day and the day after. Based on these facts, this project used the SafeGraph weekly pub visitation data in Chicago in March for the last 3 years (2019-2021), aiming to answer the following 3 questions:
Question 1
Do people choose to celebrate St. Patrick’s Day in the pubs even during the weekdays?
Question 2
Do people travel further to the pubs in Chicago and stay longer during the week of St. Patrick’s Day than other days?
Question 3
Does any other racial group increase visits to pubs during the week of St. Patrick’s Day than usual, or only white people?

Part 1: Daily bar visits in Chicago in March over a 3-year period (2019-2021)

From the plot above, We can see that the number of daily visits fluctuated in a weekly cycle, and peaks always occured on Friday and Saturday, except March 2020, as it is when the first lockdown in Chicago started and therefore the visits dropped rapidly.
To answer question 1, visits to pubs on St.Patrick’s Day did not greatly increase, and people still preferred to drink on Friday and Saturday night when they don’ have work the next day. It is worth noting that there is a significant surge on Saturday of the second week of March 2019, as the next day was St. Patrick’s Day and they didn’t need to work.

Part 2: Duration and travel distance to bars

To answer question 2, it is surprising that in the year of 2019 and 2021, the duration time at pubs during the week of St.Patrick’s Day was lower than usual, conversely, the traveling distances were at the highest level. One explanation may be that during the week of St.Patrick’s Day, more people from other places, i.e., non-locals, go to Chicago to celebrate. Compared to the locals, those people travelled further and stayed for a shorter time to save time on their return home. This might also explain the interesting reverse pattern shown in the plot below: when the travel distance increased (decreased), the duration time went down (up).
As the begining of lockdown, the third week of March 2020 showed very different patterns with much higher level of duration time and lower traveling distance. this could be possibly explained by that the visitation tracked during this week are the employees instead of customers.

Part 3: Visitation of racial groups

From the graph above, the growth of pub visits during the week of St.Patrick’s Day (week3) was mainly driven by white people. And the visitation of Asians and African American also slightly increased.

1 Like

For Part 3, which year did you analyze? And how’d you determine demo for visitors, census data?

Hi, sorry for the ambiguity here. for part 3, the timeframe is March 2021.
There is no direct demographic information in the panel, so I inferred this from some available data on cbg level: To estimate the pub visitor number of each race, I combined the ‘visitor_home_cbgs’ column from Pattern data and racial composition information from open census data, assuming the census data as ground truth in the panel.
This method is not perfect, but it did provide an indirect way to present the demographic information of the visitors.
Hope this helps.