By Joyce Chapman, Duke University Libraries
Presented on March 25, 2015 for the Trends in Analyzing Reference Data portion of the NCLA-RASS Conference Series--Trends in Reference 2015.
3. Analysis for data-driven management
0%
4%
8%
12%
16%
20%
1am
2am
3am
4am
5am
6am
7am
8am
9am
10am
11am
12pm
1pm
2pm
3pm
4pm
5pm
6pm
7pm
8pm
9pm
10pm
11pm
12am
% Research transactions by hour of the day
50%
24%
19% 18%
13%
8% 7%
12%
Research transactions by day of the
week (semester data only)
4. Problems with the data
• Estimated that up to 50% of
transactions were not being
recorded!
• Tool takes too long, too clunky
• During busy periods in particular,
staff are often unable to record
10. Suma versus LibAnalytics
• Suma has no login (saves time, but
doesn’t auto-track who recorded the
data)
• Suma can only take binary data
based on tags, does not accept text
input (cannot record what the
question/answer was)
11. Findings of pilot
• Pilot: November 2014 compared to
data from November 2013
Findings:
• 93% more transactions recorded
• 56% more research transactions
recorded
13. Obstacle to using Suma
• It is not an “out of the box” product
• It requires IT support to install and
set up (but not after initial setup)
• You will need several hours of a
system administrator’s time