From: Vincent W.J. van Gerven Oei
Pie in the Sky – Charts from the Tirana Masterplan

Today we continue our series about the Tirana 2030 Masterplan. Previous episodes included the illegal procedures surrounding its adoption, the “strategic objective” of “transformation zones,” and the wild guess of population growth.

Below I will discuss more aspects of the plan, in this case of how the underlying statistical data are (mis)represented. I will do so with two examples: water accessibility and children.

Water accessibility

For everyone who never took a statistics class, a pie chart is a form of data presentation in which numerical proportions are shown by means of the size of the “pie slice.” The proportions represented by all slices add up to hundred percent. There are quite a number of pie charts in the Tirana 2030 Masterplan, and their only job seems to be to confuse.

Take a look at the chart in figure 1.

Screenshot from 2017-01-13 08-57-46
Fig. 1. Pie chart, TR030, p. 29.

First of all, as you can see, there is no correspondence between the data presented and the actual graphic. For example, the text tells me me that 90.2% of the Tirana households has running water, while only 6.8% of the households in Zall-Bastar has it. This is shocking information. But the chart tries to tell me something else: that the percentage of the number of households with running water in the city of Tirana takes up a higher percentage of all percentages of number of households with running water in the entire municipality of Tirana than the percentage of the number of households with running water in Zall-Bastar.

If this doesn’t make sense to you, don’t worry. Because a chart that shows the relative proportions of the percentages of households with running water in the municipality of Tirana communicates something uselessly confusing: a percentage of a percentage, a meta-percentage! The chart just fills up some space trying to look important, but its information value is close to zero.

Let me show you the data in a way that allows you to actually compare things.

Fig. 2. The same data in a bar chart, sorted from high to low.
Fig. 2. The same data in a bar chart, sorted from high to low.

Figure 2 contains exactly the same data as figure 1, but now you can actually understand what is going on. First of all, we can compare the different administrative units without reading the numbers. In Tirana, the coverage of running water is highest; in Zall-Bastar, the lowest. Moreover, we can also suddenly see that there is marked difference between the availability of water in urban areas and rural areas such as Ndroq and Vaqarr.

Based on this presentation of the statistical data, you may want to conclude that the water network in these areas should have priority. But the presentation in the Tirana 2030 Masterplan, as nonsensical pie chart, makes it difficult for the reader to come to the same conclusion. Of course they talk about water supply in the text, but one image speaks more than a thousand words. There is a reason for that: improving the access to running water is not part of one of the “strategic objectives” of the plan. This is what they write about water management:

TR030, p. 272
TR030, p. 272
TR030, p. 273
TR030, p. 273

As you can see, they even had difficulty reading their own data: “Tirana currently serves it population through [water] tubes, arriving at xx% of the population” and “Tirana has currently lost xx% of its water because of leaks.” A better illustration of their illiteracy is difficult to find.

Bringing drinkable water to the inhabitants of Zall-Bastar is not a priority of this masterplan. They have however imagined the following “strategic project” instead:

The General Local Plan TR030 proposes to clean the river Lana through water machinery placed in the form of installations that will freely positioned.

The citizens will be able to interact with this simplified water filtering machinery, will be able to understand the process that is going on, while the installation that is freely positioned will turn into a device and visible object which will prove how the water changes, and how financial sources have been used. (p. 57)

This is a picture of the “freely positioned simplified water filtering device” along the Lana:

Screenshot from 2017-01-13 10-24-36
Freely positioned simplified water filtering device. TR030, p. 57


Allow me to give another example of the same problem. Take the data on couples with or without children in figures 3 and 4.

Fig. 6. Couples with children. TR030, p. 32
Fig. 3. Couples with children. TR030, p. 32
Fig. 7. Couples without children. TR030, p. 33
Fig. 4. Couples without children. TR030, p. 33

These pie charts look nearly identical, simply because in all administrative units of the Tirana municipality there are more or less the same amount of couples with and without children. In other words, these charts convey us no information additional to the text they already contain.

Let us now again look at a simpler way of presenting the same data, in figure 5:

Fig. 8. The same data presented in a bar chart
Fig. 5. The same data presented in a bar chart

Figure 5 allow us to clearly compare the two data sets offered previously in completely similar pie charts. We can also observe the very common pattern that there are relatively less couples with children in Tirana than in rural areas. But the numbers for couples living without children seem less clearly organized. What we see here is actually the complex interaction of emigration of young people, mortality rates, and natality rates as well as locality. Perhaps, these data could have been better presented divided by income group, or age. It makes little sense to do it by administrative unit, because they are all more or less the same.

But in chart after chart in the masterplan, data are cut up by administrative unit, as if that were the only logical and informative way to present the data. This is a sign of laziness and a lack of professionalism. In order to make the data speak, so that sound and intelligent policies can be built on them, it doesn’t suffice just to click “create chart” in Microsoft Excel. It requires thought about how to represent data such that they yield the maximum amount of information to the reader.

Of course, it can be justified that this specialized knowledge may be absent in the municipality. But let us remind ourselves that the government spent 47 million lekë (~€344,000) on a foreign company, Boeri Architetti, to get this right. They didn’t.

The correct presentation of data is crucial as a basis for the development of new policies. The incorrect presentation of data is only desirable when the policies are in fact not based on the data – as in the case of the Tirana 2030 Masterplan.