Page 319 -

Chapter 12
Analyzing “Spaghetti Processes”

Spaghetti processes are the counterpart of Lasagna processes. Because Spaghetti
processes are less structured, only a subset of the process mining techniques de-
scribed in this book are applicable. For instance, it makes no sense to aim at opera-
tional support activities if there is too much variability. Nevertheless, process mining
can help to realize dramatic process improvements by uncovering key problems.

12.1 Characterization of “Spaghetti Processes”
As explained in the previous chapter, there is a continuum of processes ranging
from highly structured processes (Lasagna processes) to unstructured processes
(Spaghetti processes). In this chapter, we focus on Spaghetti processes.
Figure 12.1 shows why unstructured processes are called Spaghetti processes.
Only when zooming in one can see individual activities. Figure 12.2 shows a tiny
fragment of the whole process. The fragment shows that activity “O_Bloedkweek 1”
(a particular blood test) was scheduled 412 times and 230 times followed by
“O_Bloedkweek 2” (another test). These activities are frequent. However, there are
also several activities that are executed for only one of the 2765 patients.
The process model depicted in Fig. 12.1 was obtained using the heuristic miner
with default settings. Hence, low frequent behavior has been filtered out. Neverthe-
less, the model is too difficult to comprehend. Note that this is not necessarily a
problem of the discovery algorithm. Activities are only connected if they frequently
followed one another in the event log (cf. Sect. 6.2). Hence, the complexity shown
in Fig. 12.1 reflects reality and is not caused by the discovery algorithm!
Figure 12.1 is an extreme example used to explain the characteristics of a
Spaghetti process. Given the data set, it is not surprising that the process is unstruc-
tured; the 2765 patients did not form a homogeneous group and included individuals
with very different medical problems. The process model can be simplified dramati-
cally by selecting a group of patients with similar problems. However, also for more
homogeneous groups of patients (e.g., people that had heart surgery), the resulting
process model is often Spaghetti-like.
W.M.P. van der Aalst, Process Mining, 301
DOI 10.1007/978-3-642-19345-3_12, © Springer-Verlag Berlin Heidelberg 2011