![](bg3.png)
once on different POS machines, we treat
multiple records as one meal, add the several
costs, and merged into one meal.
Step 2, we add up each student’s cost and
times monthly with breakfast, lunch and
dinner, so that a statistical chart will come into
being faster.
Step 3,we insert warning data into the table
monthly, rather than calculate each time, thus
improve access efficiency.
4 CASE PRESENTATION
4.1 Student Behavior Analysis System
We developed this system is to enable those teachers
who is responsible for student management can keep
track of dynamic information. For example, students
who have meal more than 70 times per month, but
monthly cost less than 150 Yuan RMB, etc.
4.1.1 Statistics and Query Function
One can select specific groups, such as a particular
college, a grade or a period number of students,
query and statistics the recharge records, consumer
records, eating records, computer room records,
medical records, access records, and draw the
appropriate cake Chart, histogram. As shown in
Figure 2.
Figure 2: Interface of statistical and query function.
4.1.2 Alert Function
Analyze each type of behavior, based on statistical
theory, set alerts threshold and save the results of
data analysis into the table. Such as, students who
come back dormitory too late more than 7 times a
month; students who eat more than 60 times a month
but the total cost is less than 100 Yuan RMB;
students who spend more than 300 hours a month on
internet, etc. These information need to be saved
into the question table, and take the initiative to push
a window to student management workers when
they log on, or the messages are sent to the
manager’s mobile phone.
4.2 Decision Support System
Beijing Jiaotong University have five student
canteens, a foreign student canteen and a staff
canteen. Last year, in order to build a new student
activity center, university had to place a student
canteen for demolition and reconstruction. However,
removal of the old dining hall, and rebuild a new
canteen will spend about 2 years. During the 2 years,
students will have a significant impact on crowded
dining, especially in the peak dinner time. How to
open the canteen windows scientifically, adjust the
service time of the canteens is the urgent task which
placed in front of school leaders. To this end, the
project team submitted canteen statistics data weekly
to the school leadership and logistics management.
Here are a few sample charts.
4.2.1 Comparison Charts of Each Canteen
According to the Chinese students’ consumption
data, we conducted a statistical analysis separately,
according to No.1 canteen, No.2 canteen and No.4
canteen. Our statistics date is from March 6 to 12
(except Saturday and Sunday), every five minutes as
a number of sampling points. According to statistics,
the time period as the abscissa, number of meals
eaten within five minutes as the vertical axis, make
the daily meals curve. We draw the dinner curve five
days on a chart, respectively, in different colours for
different date. As shown in Figure 3 to Figure 5.
Through analysis, We can see, from March 6 to
12, the maximum number of breakfast appears in
No.2 canteen, the peak number is 225 times every
five minutes and the peak time is 7:40. The
maximum number of lunch appears in No.1 canteen,
the peak number is 229 every five minutes and the
peak time is 12:20. The maximum number of dinner
appears in canteen No.1, the peak number is 214
every five minutes and the peak times is 18:15.
Figure 3: Analysis of dining number of canteen No.1 in
March 6 to 12.
RESEARCH AND IMPLEMENTATION OF CAMPUS CARD DATA ANALYSIS SYSTEM
517