Beginning Anomaly Detection Using

 Anomalies as Data Points

Let’s extend this same concept to a real-world application. In the following example, 

you will take a look a factory that produces screws and attempt to determine what an 

anomaly could be in this context. The factory produces massive batches of screws all 

at once, and samples from each batch are tested to ensure that a certain level of quality 

is maintained. For each sample, assume that the density and tensile strength (how 

resistant the screw is to breaking under stress) is measured.

Figure 

1-4

 is an example graph of various sample batches with the dotted lines 

representing the range of densities and tensile strengths allowed.

Figure 1-4.  Density and tensile strength in sample batches of screws

Chapter 1   What Is anomaly DeteCtIon?

---

6

The intersections of the dotted lines create several different regions containing data 

points. Of interest is the bounding box (solid lines) created from the intersection of both 

dotted lines since it contains the data points for samples deemed acceptable (Figure 

1- 5

). 

Any data point outside of that specific box will be considered anomalous.

Now that you know what points are and aren’t acceptable, let’s pick out a sample 

from a new batch of screws and check its data to see where it falls on the graph  

(Figure 

1- 6

).

Figure 1-5.  Data points are identified as good or anomaly based on their 

location

Chapter 1   What Is anomaly DeteCtIon?

---

7

The data for this sample screw falls within the acceptable range. That means that this 

batch of screws is good to use since its density and tensile strength are appropriate for 

use by the consumer. Now let’s look at a sample from the next batch of screws and check 

its data (Figure 

1-7

).

Figure 1-6.  A new data point representing the new sample screw is generated, 

with the data falling within the bounding box

Chapter 1   What Is anomaly DeteCtIon?

---

8

The data falls far outside the acceptable range. For its density, the screw has abysmal 

tensile strength and is unfit for use. Since it has been flagged as an anomaly, the factory 

can investigate the reasons for why this batch of screws turned out to be brittle. For a 

factory of considerable size, it is important to hold a high standard of quality as well 

as maintain a high volume of steady output to keep up with consumer demand. For a 

monumental task like that, automation to detect any anomalies to avoid sending out 

faulty screws is essential and has the benefit of being extremely scalable.

So far, you have explored anomalies as data points that are either out of place, in the 

case of the black swan, or unwanted, in the case of faulty screws. So what happens when 

you introduce time as a new variable?

Figure 1-7.  A new data point is generated for another sample, but it falls outside 

the bounding box

Chapter 1   What Is anomaly DeteCtIon?

---

9