In this blog post graphic methods are used to describe Serbian quarterly construction time series.

Two types of a line graph could be used to show changes in quarterly construction over time. The first displays all three time series on a single graph. Since all three series have the same base (2016 = 100) the series overlap each other.

The second type of graph displays each series on a separate panel. On such graph any persistent pattern related to trend or seasonal factors should be more clearly visible than on the first type of graph.

Figure 1 shows line graphs of quarterly construction time series on the single panel.

We can see a drop in the series levels and even negative trend in some series a year or two after the great financial crisis, i.e. 2008. While the volatilities of the *Effective hours of work* and *Value of construction work performed* series are of the same degree, the *Number of completed apartments* series shows 3-4 times greater volatility.

Because the series are overlapping, particularly since 2011, we are presenting them on a separate panels in Figure 2, so that we can clearly see distinctive patterns each of them.

A couple of features of these series stand out. First, all
three series show changes in the level. The *Effective
hours of work* series fluctuate around some constant level until 2008 and
then a negative trend until 2014. It looks like fluctuation in this series
stabilised at a lower level in the remaining period. It seems that this series quite
stable seasonal pattern due to seasonal nature of construction work. The volatility
level decreased over time.

The *Value of construction work performed* series shows rapidly increasing trend until 2008, and then fluctuation around some constant level until 2014. Since then this series kept increasing but at the slower rate than at the beginning of period. Similarly to the *Effective hours of work* series this series also shows quite stable seasonal pattern. The volatility level stayed almost at the same level during the whole period.

The *Number of completed apartments* series is an example of series fluctuating around some constant level, but with structural break that happened about 2012. Since then this series fluctuates around much lower level than before. The seasonal pattern in this series is not clearly visible. Perhaps the seasonality analysis would shed some light on this feature. The volatility level decreased in the second period after the structural break.