(1.4.5) Pattern Discovery

So far we have drilled down on the abstraction and the model. Lastly, let's dig into the pattern.

The term pattern discovery is used to collect concrete examples and to find common features or phenomena that repeatedly appear from those pieces of information. For example, when you draw a line graph with the number of everyday accesses on a web page, you notice that it is zigzag (Figure, *36). By collecting facts and visualizing it, we found a periodic pattern.

https://gyazo.com/64c527f995924d23dba46ddcda4a7a5f

Fig: Number of visits to a site

If you compare the numerical values ​​as they are written, "100000" and "110000" are very similar because only one letter is different. "100000" and "99999" are not similar at all. However, when drawing a graph, "100000" and "99999" look almost the same.

By changing the form of expression, what is focused and what is ignored changes. By expressing it in the form of a line graph, in which a small difference in number is ignored. So we can extract only an important part from concrete data. (*37)

---

Footnotes:

*36: Less people access the site in the weekends.

en.icon