Re: Anyone doing collection analysis using Millennium?
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
I'm working on a collection analysis project myself right now. I wanted
to analyze the current distribution of monographs in our main collection
by LC subclass and decade of publication and then cross-tabulate
circulation by LC subclass and decade of publication. This is been a
time-consuming task, but I think the results are going to be worth the
effort. Here's a synopsis of what I've done:
I created a snapshot of the makeup of the monograph portion of our main
collection based on LC subclass and decade of publication. I didn't use
any of the web management reports to generate this data since I couldn't
get the precise information I wanted from a canned report.
As preliminary steps, I deleted all bibliographic and item records coded
for withdrawal (suppress code=w). Since I relied on DateOne in fixed
field 008 for my publication date, I also did a little cleanup project to
fill in blank fixed fields on old catalog records when the publication
date could easily be determined from the 260 field.
First, I created a list of all item records with an item location of
"main" and Bib Lvl=m in the corresponding bibliographic record. I
deliberately excluded serials from the report for a couple reasons.
DateOne in a serial bib record reflects first year of issue, not the
publication year of any individual volume so there wasn't any easy way to
get true publication years for my report. Also, I want to apply
circulation as a parameter in our book allocation formula, so it made
sense to base my collection snapshot on monographs only so I could compare
monograph circulation to total number of monograph volumes.
Next, I created sublists of my initial review file by creating lists of
all the records in that review file subdivided by LC subclasses. I lumped
some subclasses together and separated others depending on how relevant
each subclass is to our curriculum.
Once I had a review file of all item records for a particular LC subclass,
I created additional lists of all the records in that review file by
decade of publication using DateOne of the fixed field. My decade
breakdowns were <1950, 1950-1959, 1960-1969, 1970-1979, 1980-1989,
1990-1999, and > or = 2000. As I determined how many volumes we had in
each subclass and decade of publication, I jotted the numbers down in a
handwritten table that I later transferred to an Excel spreadsheet.
Creating all of these review files took me about 5 days with
After getting my collection data, I started on circulation data. We
migrated to Millennium and moved into a new building in 2001, and I
thought circulation data since that time would be a useful time period. I
thought I could use the TotalCheckouts field in the item record, but I
discovered that we had migrated our old circulation data from our previous
ILS system into this field, so I had up to 16 years of data for items that
we had owned since 1990 when the first ILS was implemented. That time
span seemed too long to reflect current usage since book circulation prior
to the web and the growth of electronic resources was undoubtedly
different than book usage in recent years. Since we have only moved
YTDCirc data to the LYCirc field once during our four years on Millennium,
I decided to use a combination of those two fields to get my circulation
data. Anyone trying to create a similar report will want to make sure
they know what time span is covered by these three fixed fields in order
to pick an appropriate field for collecting data.
Using my initial review file again, I created a new list of items in that
review file that had YTDCirc >0 or LYCirc >0.
Next, I sorted the new review file by call number and then exported the
following data: call #, title, 008DateOne, YTDCirc, and LYCirc. I
imported this data into Excel. I had to perform a little clean up of data
that didn't parse correctly (usually because of punctuation in the bib
record) and got put into the wrong columns, but most information imported
cleanly. If a bib record had alternate call numbers, I got duplicate
items in my spreadsheet. These duplicates could easily be identified by
the presence of quotation marks in the call number column of my
spreadsheet, so I used the search feature of Excel to find these and
eliminate the duplicates.
I had Excel add the YTDCirc column to the LYCirc column, and then I had
total checkouts for each item since we went live with Millennium
circulation in July 2001.
Next, I split my spreadsheet up by the major LC subclasses I'd already
determined when compiling collection data. In my column for 008DateOne, I
applied the A-Z function to get all the data for each subclass grouping in
chronological order by publication date. I had just a few items with
unknown publication dates, and I lumped them in the <1950 publication
group. I then split my data within each subclass by decade of
publication. I figured how many volumes within each decade had circulated
at least once, and I used the autosum feature on my column of circulation
data to find out how many total checkouts there had been of books in each
subclass from each decade of publication. It took another week or so to
finish all these tabulations. As I figured all these numbers, I jotted
them down in a handwritten table and later input them in an Excel
After I had all my collection data and my circulation data, I typed
everything into an Excel worksheet. I have LC subclasses running in rows.
Under each decade of publication going across the spreadsheet, I have a
series of columns for Total Volumes published within that decade, # of
Items with Checkouts, Items with Checkouts divided by Total Volumes, Total
Checkouts within the decade, Total Checkouts divided by # of Volumes, and
% of Total Circulation within the Decade. The concluding columns show
summary data for all monographs in our main collection.
This spreadsheet was ungainly and had to be printed in tiny font to
squeeze it onto two pages, so I separated out each decade and the summary
data and created separate worksheets for each time period. I added a few
additional columns to these worksheets, so under each decade of
publication I have columns for: Total Volumes, % of All Volumes in
Collection Published in Time Period, % of Total Book Collection, # of
Items with Checkouts, # of Items with Checkouts divided by Total Volumes,
Total Checkouts, Total Checkouts divided by Volumes, % of Total
Circulation within Time Period, and % of Total Overall Circulation from
July 2001 to Sept. 2005 (when I collected my data).
I am now happily creating charts and graphs to illustrate the circulation
trends in our book collection and to identify what portions of our
collection circulate and what doesn't. I spent yesterday making pie
charts of each LC subclass to show the makeup of each subject collection
by decade of publication and the circulation of the same segment of the
collection by decade of publication. For most subject areas, the two pies
don't bear much resemblance to each other. We have a huge glut of
materials published in the 1960s and 1970s, but this big segment of the
collection doesn't account for all that much circulation. As expected,
newer books tend to account for the greatest amount of circulation, but
this varies greatly by discipline. Circulation of literature and history
tends to be more evenly distributed across publication periods than other
disciplines, while more than 2/3 of the circulation of math and computer
science is from books published in the 2000s. Subclass QA is the one area
of our collection with more checkouts since 2001 than the total number of
volumes in the subclass.
I've just started to try to depict graphically what hasn't circulated in
each discipline with the thought that we could pinpoint which areas of the
collection are most in need of weeding based solely on the criterion of
rates of noncirculation. Not surprisingly, books published in the 1960s
and 1970s are taking the lead in my initial charts of noncirculation for
the collection as a whole.
This has been a complicated process to try to describe in words. I'd be
happy to answer any questions.
Head of Library Systems & Operations
FSU Library for Information, Technology and Education (FLITE)
Ferris State University
1010 Campus Dr.
Big Rapids, MI 49307
"Daught, Gary F." <GFDaught at milligan dot edu>
Sent by: innopac-bounces at innopacusers dot org
10/13/2005 09:52 AM
Please respond to
IUG INNOPAC List <innopac at innopacusers dot org>
<innopac at innopacusers dot org>
Anyone doing collection analysis using Millennium?
Greetings. I am new to the list and relatively new in my position as
Reference and Collection Development Librarian at Milligan College, TN.
Our library uses Innovative Interfaces Millennium (Silver). I have
really only just begun to appreciate the capabilities of this software.
I have recently received a mandate from our Library Committee to develop
strategies and procedures to analyze our collection's strengths and
weaknesses relative to our college curriculum. It strikes me that the
Create Lists function would be a key tool for me to begin to look at
However, before I attempt to 'reinvent the wheel' on this, I would
solicit feedback from the list. What are you doing in the area of
collection analysis using Millennium? What specific procedures have you
developed, or what pre-configured resources are you aware of to assist
you in this task? Can you direct me to any documentation or tutorials?
Thanks for your kind assistance!
Gary F. Daught
Reference and Collection Development Librarian
P. H. Welshimer Memorial Library
Milligan College, TN 37682
gfdaught at milligan dot edu
--- StripMime Report -- processed MIME parts ---
text/plain (text body -- kept)