San Francisco Public Library Data Usage

About this dataset: Containing anonymized data usage by over 420,000 patrons, this dataset was sourced through Kaggle here.

Data Overview

Before diving too deeply into the data, I wanted to examine things from a broad perspective. Using Excel pivot tables, I created the following visualizations regarding checkouts, patron registration year, and the correspondence between these two entities.

1. Rise of Non-Users

To begin establishing the baseline data to examine the decline in use, I first wanted to get a bird’s eye view of the number of users that registered in the years between 2003 – 2016 (the span of this dataset). I first created a query that pulled together all individual users and their registration year. I then created an expression to count the number of registrations for each year and grouped users by registration year. I then created a report to give a concise overview of the number of users registered each year. 

I exported the library use query to Excel, where I created a pivot chart summarizing the number of new patrons and the number of checkouts. I then created a 3D comparative bar chart that would illustrate how disproportionate the relationship between these two groups is – it is clear in looking at the chart below (2. Customer Loyalty), that these never came close to keeping pace with one another.

Here, it seems the 10 – 19 age group has the highest number of checkouts, though it is a close number as compared with the 45 – 54 and 65 – 74 age ranges.

To expand further into this line of inquiry, I decided to look at checkout numbers, as these give a clearer picture of what is driving circulation. I was interested in exactly how the circulation numbers changed over time, and if the same age groups were more active or if there were some potential changes, maybe due to the expansion of programming aimed at a specific group or an effort on the library’s part to be more active in outreach.

As is visible in this heatmap, the proximity between multiple age groups is clear, though some very interesting information is visible here. The proportional growth of these groups in the last year of available data is staggering to say the least. The age groups 65 – 74 and 0 to 9 years both increased tenfold between 2015 and 2016. That kind of growth falls well in line with the fact that there was a universally significant jump in membership in 2016. This aspect of the user base will be discussed in a later section regarding the branch expansion. 

Local Branch Data and Analysis

Though this is fairly straightforward, there are a few things in the above chart worth noting. Remembering that this data represents library use from 2003 to 2016, the top three libraries in library circulation are the Main Library (21.14%) and Chinatown (9.04%), both branches opened in 1996. This means that both of these branches were well-established and firmly rooted in the SFPL system before the start of this dataset. The Richmond branch (7.69%) opened in 2009, yet has surpassed all but the Chinatown and Main Library branches in this dataset (2003 – 2016).

Branch Expansion Data

When looking at the above data regarding circulation and renewals, there is quite a spike in 2016, which begins its upward trajectory in 2012. Looking into the history of SFPL, one notes that there was a rapid expansion and addition of several new branches during this interval. This certainly helps explain how circulation numbers doubled between 2011 and 2012, and by 2016 circulation numbers were almost 24 times greater than 2011