Analyzing World Health Statistics 2005 using TreeMap
and Spotfire.
Abhinav Gupta
1.
Introduction- The
Dataset
The World Health Organization
(WHO) collects and summarizes a wide range of quantitative data from a variety
of health domains through country offices, regional offices and headquarters
departments. These data are used internally by WHO for estimation, advocacy,
policy development and evaluation. They are also widely disseminated in formal
publications and through more informal mechanisms, both in electronic and
printed format.
We use one of the health statistics data for the year 2005 to find some interesting facts about the health status, services and system of the 192 nations of the world. The member nations of the UN have been divided in six WHO regions (See Figure 2). These six regions are:
·
African Region (AFR)
·
South East Asian Region (SEAR)
·
Region of the
·
European Region (EUR)
·
· Western Pacific Region (WPR)
The data collected for member nations includes information on the demographics, morbidity, mortality, health system, risk factor and health services. We analyze the following data(23) for each member nation:
1.1 Missing Data / Values
Many member nations fail to
collect several data due to lack of resources or the public policy specific to
the nation. In these cases, we have chosen to filter those member nations from
our analysis whose data is either not available or not applicable for
comparison.
We used -1 value for all missing data and filtered the data having negative value while analyzing the data with the help of treemap and spotfire.
2. Visualization Tools
2.1 Treemap
Treemap is a space-constrained visualization of hierarchical structures. It is very effective in showing attributes of leaf nodes using size and color coding. Treemap enables users to compare nodes and sub-trees even at varying depth in the tree, and help them spot patterns and exceptions
2.2 Spotfire
Spotfire Decision Site offers an interactive, visual approach to data analysis that empowers individuals to quickly and easily see trends, patterns, outliers and unanticipated relationships in data. For our case the legend for the spotfire across the figures is shown in Figure 1.

Figure 1: The Legend
for Spotfire Figures.
3.
The Demographics
We analyzed the demographics data
using the Treemap software. This helped us analyze the regional distribution of
Population. As it is evident from the figure 2, almost half of the world's
population resides in SEAR and WPR region which includes

Figure 2: The
Population Distribution and the growth rate. The size of the rectangle is proportional
to the population of the
country and the color
depends on growth rate. The red shows a negative growth rate and green shows a
positive growth-rate.
The per-capita income fails to
show any regional coherence which could have been expected from the results
above. Only a very few nations in every WHO region seem to have higher
per-capita incomes. While most of the

Figure 3: The
Population Distribution and the per-capita income. The size of the rectangle is
proportional to the population of the
country and the color
depends on per-capita income. The red shows a lower per-capita income and green
shows a higher per-capita income
4.
Money Matters
4.1
Money is the key to better living conditions
We tried to understand the correlation between the per-capita income and living conditions. We plotted the sanitation access to per-capita income using spot-fire (See Figures 4,5). It was interesting to see that sanitation conditions improve with log of per-capita income.

Figure 4: The Sanitation Access improves with per-capita
income but logarithmically.

Figure 5: The Sanitation Access improves with per-capita
income (In a Treemap). Rectangle size
being per-capita
income and bright green implies more easy access to sanitation.
4.2
Money leads to better health statistics.
Improved living conditions also lead to better health status among nations. This can be seen by plotting percentage of children under 5 stunted for age against log of per-capita income (See Figures 6, 7). Again a linear correlation can be easily seen.
Figure 6: Children Under-5 are less stunted if the per
capita income of the nation is high.

Figure 7: Children Under-5 are less stunted if the per
capita income of the nation is high (In
a Treemap). Rectangle size
being per-capita
income and bright green implies lesser children stunted for age
4.3
But Money brings vices.
People earning more tend to drink
more alcohol. The nations with higher per-capita income have higher alchol
consumption rates then nations with lower per-capita income. It is interesting
to see the black dots in Figure 8 and EMR region in the Treemap (Figure 9).
Here, the alchol consumption seems to be generally less than other nations.

Figure 8: Earn More Drink More. Nations with higher per-capita
income also have higher alcohol consumptions.

Figure 9: Earn More Drink More. Nations with higher per-capita
income also have higher alcohol consumptions (In a Treemap). Rectangle
size being per-capita
income and bright green implies more alcohol consumption
5.
Money buys LIFE everywhere except
If money can be so important, can money buy life? We were
particularly interested in looking at correlations between life of males (similar
observations in life of females) and the per-capita income. It was surprising
to see that while money really does buy life everywhere in the world, it fails
to do so in the land of

Figure 10: Money Buys LIFE

Figure 11: But not in
6.
Are Health Services and System in
We compared per-capita health
expenditures with per-capita incomes for 192 member nations. Most of the
nations tend to follow the linear pattern where Nations with higher per-capita
income tend to have higher expenditures on health services and systems.
However,

Figure 12:
7.
Region of
We compared the number of physicians to number of nurses per 10,000 people for 192 member nations. The AFR, EMR and SEAR regions are short of nurses and physicians. However, when we only plotted EUR and AMR region nations we found that AMR nations face a sever Nurse shortage as compared to EUR region. European region seems to have a good Nurse/Physician ratio (See Figure 13).

Figure 13: AMR region is facing a shortage of nurses.
8.
TB: Effecting Life Expectancy across Africa and
We plotted number of TB Cases per 10000 people and the life expectancy for 192 member nations. We found that for AFR and SEAR Life Expectancy is inversely proportional to number of TB cases reported in these regions (See Figure 14). This affect is particularly strong in the regions of AFR and SEAR. While the EUR and AMR regions seem to be less affected by it.

Figure 14: TB a major problem in land of Africa and

Figure 15: TB has not such severe effect in AMR and EUR.
9.
Tools
Both the Treemap and Spotfire are
easy to use softwares. Spotfire even suggests some graphs which have some
interesting observations. However, both seem to require very specific inputs.
The treemap software fails to take in the inputs with extra empty lines. There
is no special way to handle missing data in the either so I had to use -1
filter in both the softwares.
Spotfire has lot of tools for data analysis like scatter plots, profile plots, histograms, pie charts. Treemap seems a very nice concept to understand hierarchal relationships in data.
10.
Conclusion
While the study suggests that
money seems to improve life conditions it seems that money is not properly
being used in major african nations. The study also suggests that