Visualize Better Library Data Presentations

Libraries capture data of all sorts.  Sometimes we need to mine it but whether it’s visits to the catalog or transactions in one of our subscriptions, the data is lying around.  I find presentations of tabular data to be about the least useful way to see the story that the data is supposed to tell.  I have been trying to play around with a variety of different methods of having the data tell something more useful than just its contents.

Heat Maps in Microsoft Excel

Library electronic subscription usage data.  The left hand side shows the data - searches in libraries by date - with the libraries in alphabetical order.  The right shows the same data, but sorted on total number of searches.

Library electronic subscription usage data. The left hand side shows the data – searches in libraries by date – with the libraries in alphabetical order. The right shows the same data, but sorted on total number of searches.

This was one of the first attempts I made.  Using Microsoft Excel, you can apply a heat map effect to the data using conditional formatting.  As you can see, it wasn’t until I’d thought about it a bit longer that I realized that, by reordering the data using a column that totaled the other data (but wasn’t itself part of the heat map), I had a much more interesting representation of the data.  It is easy to see that, in some of the library locations, very few searches were occurring.  It is hard to pick these out when they are mixed up with the higher activity library branches.

Other Microsoft Excel Visuals

Library collections frequently represent a funnel, with a wider top (current) and a narrow tip (historic).  As a law library, we have a bulge in the year Canada was created.

Library collections frequently represent a funnel, with a wider top (current) and a narrow tip (historic). As a law library, we have a bulge in the year Canada was created.

I have frequently referred to law library collections as funnel-shaped.  Most library collections share this shape – broader at the top, with current materials, and tapering off over time as only the classics or most important print materials are preserved.  Normally, the data in a Microsoft Excel chart hugs one axis.  But you can make it look like a funnel – and the idea is from someone who did a sales funnel – by creating an extra row of data that is transparent.

There is also the traditional colored Excel chart.  I use these frequently, sometimes using an accent color to highlight the important element.  In this one, I used a standard color format for the chart and then customized just the color applied to looseleaf publications.  This binder format is increasing at between 10-15% a year in our budget and forcing out other parts of our collection.  You can get the cancellations to hang off the bottom by giving them a negative number (so $100 in cancellations is represented in Excel as -$100).

Looseleaf publications are consuming increasing amounts of the library's budget, at the expense of other parts of the collection.

Looseleaf publications are consuming increasing amounts of the library’s budget, at the expense of other parts of the collection.

Another chart I’ve created uses the basic chart again but I’ve overlaid some text boxes on it to bring in a second set of information.  I don’t think it’s too cluttered and it gives a bit more of an infographic representation to our operations data.  The Windows Wingdings font has a small person symbol, so I used different colors to represent types of roles with a shadowy color to represent the eliminated positions.

Our library's operational data, showing funding and staffing trends over 12 years.

Our library’s operational data, showing funding and staffing trends over 12 years.

Big Data

Most of my data work is within Microsoft Excel, as you can see.  I was reading a book recently on big data that mentioned a site called BigML.  It enables you to upload large amounts of data and will parse it for you, enabling very quick visualizations of data.  In this case, I took two years worth of searches – about 35,000 rows of data – and put it into BigML.

This use of BigML’s Dynamic Scatterplot shows usage across a variety of locations by the day of the week – Monday is at the bottom, Sunday is at the top.  Large usage tends to correlate to size of library or how urban v. rural it is.

Usage data by the day of the week.  This could be used to contrast usage rates of different services too.

Usage data by the day of the week. This could be used to contrast usage rates of different services too.

You can see that most locations have pretty consistent usage across the weekdays of Monday through Friday but only some have weekend usage.  Even if you work in a library without branches, you could use data of different types to contrast days of the week that certain services or resource were used more or less.  It could help you in redeploying staff or money depending on how the resources are used.

Leave a Reply

Your email address will not be published. Required fields are marked *