Library Data

Collecting and analyzing data about libraries and their future .
The shape of public libraries
Here I picked library visits, but if you pick pretty much any feature of public libraries and plot out a frequency histogram, it will look pretty much just like this.  What’s surprising about this data? It has a very strong positive skew, which in the case means that for the most part, public libraries are small.  There is of course a very long tail of large libraries (in fact I trimmed off some of the largest just because it became so hard to see the rest). 
Data Source: IMLS

The shape of public libraries


Here I picked library visits, but if you pick pretty much any feature of public libraries and plot out a frequency histogram, it will look pretty much just like this.  What’s surprising about this data? It has a very strong positive skew, which in the case means that for the most part, public libraries are small.  There is of course a very long tail of large libraries (in fact I trimmed off some of the largest just because it became so hard to see the rest). 

Data Source: IMLS

Public Library Closings in Decline
Extracting this information from the IMLS raw data was harder than expected.  Not until 2008 and 2009 is there any indicator that a library has ‘closed’.  After researching a bit I’ve decided the best estimation is described in this paper [pdf]. The idea is essentially: look for missing entries between years, and use that as a best guess for closing.  To test this out I also looked at how many entries remained missing after 2,3,4 and 5 years, as well as hand checking the results.  My conclusion is that this is a reasonably reliable method to determine.
Now that the issue of ‘how did we get this data’ is out of the way, the results are fascinating.  Confirming what Walt Crawford had mentioned in a post not long ago the state of public library closings is not actually as bleak as it seems.  From the data we have it even appears as though public library closing are actually declining over time!
I’ve definitely heard a lot of talk about public library closings, but, anecdotally, whenever I would investigate further I would frequently find that at the last minute plans to close were cancelled. The results above lead me to believe people threaten to close public libraries much more frequently than they actually do.  

Data Source: IMLS

Public Library Closings in Decline


Extracting this information from the IMLS raw data was harder than expected.  Not until 2008 and 2009 is there any indicator that a library has ‘closed’.  After researching a bit I’ve decided the best estimation is described in this paper [pdf]. The idea is essentially: look for missing entries between years, and use that as a best guess for closing.  To test this out I also looked at how many entries remained missing after 2,3,4 and 5 years, as well as hand checking the results.  My conclusion is that this is a reasonably reliable method to determine.

Now that the issue of ‘how did we get this data’ is out of the way, the results are fascinating.  Confirming what Walt Crawford had mentioned in a post not long ago the state of public library closings is not actually as bleak as it seems.  From the data we have it even appears as though public library closing are actually declining over time!

I’ve definitely heard a lot of talk about public library closings, but, anecdotally, whenever I would investigate further I would frequently find that at the last minute plans to close were cancelled. The results above lead me to believe people threaten to close public libraries much more frequently than they actually do.  

Data Source: IMLS

Note: This is just a first stab at extracting information from IMLS reports, which it turns out is not as easy as I had originally hoped ;) see next post for more accurate findings.

Preliminary closing data
I recently came across a few blog posts that discussed having some difficulty finding data on public library closures.  So I decided to dive into the raw data files from the IMLS and see what I could find.  Here are the numbers for 2008 and 2009 the only years that specifically record closures.  For a sense of scale each year we see about 1/2% of public libraries close, and half of that are branch libraries.  
I have a few ideas on how I can expand this (look for libraries that don’t repeat in the next years data). I’d really like the expand on this is data a bit more, and I have a few interesting projects I’d like to try out as well.  Sometime soon expect data for years from 1998-2009 if not further back.
As an added bonus I’ve also posted the source for the code I’ve been working on tonight in order to pull this data together. Feel free to play on your own!
Data Source: IMLS

Note: This is just a first stab at extracting information from IMLS reports, which it turns out is not as easy as I had originally hoped ;) see next post for more accurate findings.


Preliminary closing data

I recently came across a few blog posts that discussed having some difficulty finding data on public library closures.  So I decided to dive into the raw data files from the IMLS and see what I could find.  Here are the numbers for 2008 and 2009 the only years that specifically record closures.  For a sense of scale each year we see about 1/2% of public libraries close, and half of that are branch libraries.  

I have a few ideas on how I can expand this (look for libraries that don’t repeat in the next years data). I’d really like the expand on this is data a bit more, and I have a few interesting projects I’d like to try out as well.  Sometime soon expect data for years from 1998-2009 if not further back.

As an added bonus I’ve also posted the source for the code I’ve been working on tonight in order to pull this data together. Feel free to play on your own!

Data Source: IMLS

Surplus of MLIS degrees (US)

This has been one of the more popular topics I’ve discussed so I’ve decided to take a closer look at the data. Earlier I pointed out that there has been a tremendous increase in the number of MLIS degrees conferred over the years.  In these two charts we’re taking 2 different looks at the same data.  

The rate that the profession grows is a huge factor in considering whether or not we should be concerned about rising numbers of MLIS grads.  In the first chart I’ve taken this into account.  I’ve taken the change in librarians from the previous year and subtracted that from the number of new MLIS degrees conferred.  If, for example, the change in the number of librarians employed from one year to the next is roughly equal to the number of MLIS grads, then we can expect everyone to find a job.  If the number of new librarian position is greater than the number of new grads, the we can expect to run out of enough trained people.  And finally if growth smaller is than the number of new grads we end up with a surplus of MLIS degrees.  Blue bars in the first chart means that the number of librarians in the US did not grow to meet the number of new MLIS graduates.

In the second chart we can see the cumulative effect. Ideally we’d want that line to be roughly flat.  Clearly this is not the case, in this chart at at 2009 we can see that 46,648 more MLIS degrees have been generated than jobs created.

The one factor I haven’t taken into account is retirements, growth creates jobs as well as retirements.  I’m still working on creating some models for this, but even anecdotally we can guess there may be trouble.   

46,648 librarians would have had to retire from 2000-2009 to make enough jobs.  In 2000 there were 139,460 librarians in the US.  This means roughly 1/3 of all librarians working in 2000 would have had to retire in order for there be enough jobs to meet the needs of new graduates with MLIS degrees.  However retirement data, and even good projections, are hard to come by.  I’ll do my best to show a couple of best and more realistic scenarios.

Data sources:

Bureau of Labor Statitics

Digest of Education Statistics (2010)

The most important public library statistic?

In other posts we’ve looked at various changes in public library services over the last 10 years, trying to see what the reality is behind anecdotal fears that “public libraries are doomed!!!”.  We’ve look at circulation, visitation, reference and library programs, and the results are mixed as far as indicating the long term future of library services. 

But when you think about the future the reality is that funding is going to be tremendously important to the survival of public libraries.  We’ve already seen that employment has been steady, and the ratios of types of employees consistent.  Now we see that library budgets have been growing, even on a per capita basis. We do see this trend flatten out around 2008, so the big question on the future of library funding is going to be: as the recession eases up, do libraries start to get more per capita funding?

After looking at public library data for a week I think it’s fair to say that it is too early for  a lot of doom and gloom regarding public libraries.  At the same time, there are traditional libraries services that are in rapid decline (reference), it is important the public libraries be open to change and grow to meet their users changing needs.

Data (and Chart) Source: IMLS Public Survey Fiscal Year 2009

Public library programming

Before we saw the reference is definitely on a decline in public libraries.  Here we see some upwards trends in library programming.  On a per capita (or per 5000 in this case) we can see that library programming in on the rise.  However, one again, when we normalize we see that for general library programming things have actually remained steady, and for children’s programming we actually see a decline in attendance.

A note on the children’s programming, it could be possible that less people are going to children’s programming or the data could also be argued to show that the number of programs offered is not keeping up with the number of visitors.  It’s impossible from the data alone to know the causal relationship: are children’s programs growing at a slower rate because there’s low attendance, or is attendance caped because there aren’t enough programs. It would certainly be a worthwhile experiment to observe the response to having more children’s programs.

Over all what we’ve seen with public libraries is that we have people visiting public libraries more frequently, using reference less, and everything else about the same.

Data (and Chart) Source: IMLS Public Survey Fiscal Year 2009

Declining reference

One area where public libraries really are changing rather dramatically is reference.  In the per capita results we see a general decline in reference transactions, and when we look at per visitation the trend is really obvious.  

So far what we’ve seen in our data about public library users is that we have more people coming to public libraries, and they’re talking out about the same number of materials and asking a lot fewer reference questions.

Data (and Chart) Source: IMLS Public Survey Fiscal Year 2009

Public library circulation

When only looking at per capita changes, we see a steady rise in circulation, but what’s interesting is that when we look at circulation per visit, it’s actually relatively flat.  The meaning of this is more people are visiting the library, but they aren’t checking out more books (nor are they checking out less).  What I find really fascinating about this is actually what it means compared to a similar set of data for academic libraries.  While it does seem that users of academic libraries are having behavior change, it’s not clear that this same change is happening to the users of public libraries (at least it’s not reflected in the circ data).  

From the perpective of “What is the future of public libraries?” this data is at the very least a statement that current library services should grow at a steady rate, and that it’s anything but certain that their future is limited.  

Data (and Chart) Source: IMLS Public Survey Fiscal Year 2009

The future of public libraries

It’s not uncommon to hear a lot of concern about public libraries coming from librarians and paraprofessionals working, here’s a really interesting example of this from meta-filter.  Anecdotally I’ve heard very little concern from academic libraries, but looking at the data has made me less certain of even the relatively near term future.  So the most interested data question for public libraries is “are these fears justified?”.  Earlier I showed some data on public libraries FTE employment.  Except for hints of decline at the end of the data, things look relatively good.

Above are 2 charts taken straight from the IMLS Public Survey Fiscal Year 2009, and they both tell pretty different stories. The first shows that only 1 state in the US has seen an increase in the per capita number of libraries, nearly half have seen a decrease. The other shows that visitation per capita is up! I’m sure there are many who work in public libraries that would anecdotally confirm this, less libraries are more busy.

Data Source: IMLS Public Survey Fiscal Year 2009

Continuing scale - publishers vs public libraries
At first this comparison makes libraries seem closer in scale to publishers… until you look at the labels.  What we’re looking at is the total volume sold by publishers in 2009 compared to the total number of volumes in the collections of public libraries in 2009. Put perhaps more clearly: In 2009 US Publishers sold books numbering the collections of all public libraries 3 times over.

Data sources: 
Association of American Publishers
Public Libraries Survey 2009

Continuing scale - publishers vs public libraries

At first this comparison makes libraries seem closer in scale to publishers… until you look at the labels.  What we’re looking at is the total volume sold by publishers in 2009 compared to the total number of volumes in the collections of public libraries in 2009. Put perhaps more clearly: In 2009 US Publishers sold books numbering the collections of all public libraries 3 times over.

Data sources: 

Association of American Publishers

Public Libraries Survey 2009