Introduction to Plots and Maps
Bob Molyneux
What we have here is a sample of graphic representations of the behavior of library variables over time. They resulted from initial programs to examine the longitudinal data I have been working on and to construct general programs to present the data graphically. Often times such graphic representations of data aid understanding of phenomena. Suggestions of other variables to examine are welcome. A goal is to automate as much of this process as possible but for now, there will be a bit of experimentation and programming but, if you want something plotted, I will do what I can. Yes, groups of states can be done, as can regions.
Data from three different series are used here:
- State Summary/State Characteristics data (1992-2001) from NCES as recompiled and documented elsewhere on this site. This series uses imputed data in an attempt to arrive at a best guess of state summary data about public libraries given that not all libraries fill out the survey forms and some fill the forms out with information that is dubious or incorrect. NCES uses various techniques to create or modify data in an attempt to overcome the vagaries of collecting the data for the purpose of arriving at accurate assessments of the condition of libraries nationally and in the states.
- The State Library Agencies (StLA) data (1994-2001) also compiled by NCES. These data are currently not available from NCES but a subset of this dataset has been reorganized in the same manner as the State Summary data, that is, into a longitudinal file. Data from this subset of the StLA data are plotted here. This series of data is a good bit more complicated than the others that have been recompiled so documentation of this longitudinal series will take some time.
- Economic data from a number of public sources have also been assembled. These data are documented where appropriate. As of this writing, these data have come from the US Census Bureau's State Local Governments Finance site, the Bureau of Economic Affairs (see links under "Regional"), and the Bureau of Labor Statistics.
The Public Library Survey actually started in 1987 when data from 19 states were assembled by NCES and subsequently published. This series now comprises a detailed annual survey of the libraries in each of the states and a few other areas. These data have also been assembled in a longitudinal file and documented elsewhere on the NCLIS site. So far, this is the largest and most comprehensive of these data files but, although they give us more years to analyze and a rich series of data, these data have comparability problems from year to year. The programs that generate the plots could be run against these public library data but these comparability problems will inhibit our ability to draw reasonable conclusions. These problems can be fixed with work.
The sample plots and maps presented here are evolutionary. Those done later have more information on them such as citations to the source and the like. The earlier ones will reflect the fact that they were developmental.
Graphic representation of data has enjoyed a resurgence particularly since the work of JW Tukey who gave us the term "Exploratory Data Analysis" and techniques have been discussed by numerous authors, probably best known is Edward R. Tufte's The Visual Display of Quantitative Information. For many purposes, such as exploring a new set of data, graphic techniques are useful; they are also useful in planning and budgeting. However, they can also deceive the eye so caution is in order. In the expenditures graphs, I will give some examples.
These plots and maps were done using SAS and were converted from SAS output to jpeg format. I am not certain that this procedure is the right one for presentation graphics but it is a start and it works on the Web. Meanwhile, printing the maps on a color printer gives varying results.
Links to the Plots and Maps
Return to NCLIS Homepage