Annex C: Using a computer for demographic projections and map-making

C.1 Using a computer for demographic projections: wanting to know what lies ahead

One of the things a community may wish to know is what the local population will be in 10, 20 or even 50 years. How many school-aged children will we have when our children are young parents? How many people will be of working age, or old age (over 65)? What will happen if the new migrants keep coming in?

These questions can be answered easily and precisely using demographic projection programs that have been designed for use with small computers. With such a program, the user can obtain a projection of population by age and sex for 5 to 50 years after the base or beginning year. The user decides on the base year and the number of years for the projection, and can also make a few assumptions concerning future trends for birth and death rates. It should be stated at the outset that a projection is only as good as the assumptions upon which it is based. However, the advantage of using computers is that the assumptions can easily be changed from one projection to the next to see the implications of different patterns of births or deaths on future population size and age distribution.

Projections can be obtained on urban and rural populations separately, and on populations of specific areas of interest, such as a local community. Migration can be taken into account, with the user deciding on either current migration rates, or different assumptions of flows of migrants. Some of the programs can provide information on the future impact of AIDS, if the user can provide basic information on current infection rates and assumed future rates. In these ways, the computer programs can be a good tool for helping communities see the future implications of population interventions.

Inputs and outputs

The basic inputs for running computer projections are the key population indicators discussed in Annex A:

• • population size, including the date of the information, and the distribution by age and gender, i.e., numbers of males and females for five-year age intervals: 0-4, 5-9,… 80+. Such data are usually available from national population censuses down to province or district levels, and sometimes even for small communities. The information could also come from a participatory community census.
• • total fertility rate (TFR): one can use the UN-projected TFR, or choose a different one. If available, one can use a full set of age-specific fertility rates.
• • mortality data, either as age-specific death rates, or choosing a life table and entering simpler assumptions of life expectancy at birth for men and women, again by five-year intervals.
• • migration rates: best if age-specific migration rates, usually obtained from the participatory assessment of the actual community. This is important for small communities, which can be overwhelmed by relatively small amounts of immigration, or badly weakened by even small amounts of out-migration.

The basic outputs are tables of numerical data or graphs showing total population, or numbers for specific age groups for the selected time into the future. These programs can also show projected numbers of births or deaths, or summary demographic statistics and age-sex pyramids.

By inputting different data or assumptions, it is possible to compare alternative future outcomes in the community. This is especially useful because it shows clearly and dramatically what are the future implications of different real conditions or interventions. For example, one can see the future implications of conditions that might be controlled with specific programs, such as death, fertility and migration rates. With appropriate consideration given to local cultural sensitivities, one can generate tables and graphs of future population scenarios, and use these to stimulate group discussion.

An illustration: the future implications of changing fertility

The following illustration shows how such a computer program could be used. For example, the issue of children, family size or family planning might come up in a participatory appraisal. One of the questions that commonly arises is 'What will the future hold if things continue as they are, if we continue to have the same numbers of children or the same family size?'

The following is an example to illustrate both what a projection program can provide, and what are some of the important lessons that can be learned from such an exercise (see Annex A for definitions of terms).

Take two hypothetical (make-believe) communities, with demographic characteristics that are quite common for many developing countries. They both begin in 1970 with 2,100 people, and slightly more females than males. They have age-sex distributions common among communities with high fertility and recent rapid declines in mortality. They are both at relatively high total fertility rates (TFR 7.00). In mortality, Community 2 begins with slightly lower rates and a moderate advantage: crude death rates for Community 1 and 2 are respectively 26 vs. 16, and infant mortality rates are 167 vs. 127.

Then we assume that Community 1 shows very slow declines in both mortality and fertility. Infant mortality declines slowly, remaining above 100 until after 2005, and life expectancy for males and females together increases only slowly from 38 to 58. The total fertility rate declines only gradually down to 5 by the year 2020. This is, in fact, a scenario close to that projected for a typical poor rural community without any specific local interventions.

We assume that Community 2 shows the kind of rapid decline in mortality and fertility that Thailand showed after 1965. Infant mortality goes down from 127 to 37 in 25 years, and life expectancy for males and females together rises in that same period from 49 to 73. The total fertility rate declines from 7 to replacement level in 25 years (1970-95). Not only Thailand, but South Korea, China and Taiwan had similar and even more rapid changes in mortality and fertility.

It may be an important lesson to note that these are not wild assumptions either for a small community or for a very large nation. They reflect what can be done when a government (or a small community) is committed to, and effective in, delivering primary health care and family planning services to its people. It also shows what is happening in so many countries today, where primary health care does not reach most of the people, especially the rural poor.

When we have made the basic assumptions of mortality and fertility declines, the computer program can project numbers of people by age and sex, and numbers of births and deaths for every five-year period for the 50 years 1970-2020. Table C.1 shows just a portion of one of the total population tables. One can see exactly where the decline of fertility overtakes the decline of mortality to bring smaller numbers. In 1990, 20 years after the starting date, Community 2 has 1 person less than Community 1: 3,460 vs. 3,461. From that point, however, the difference becomes very dramatic. By 2000, Community 1 already has 600 more (about 15 percent). By 2010, Community 1 has 1,685, or 38 percent more; and by 2020 Community 1 is 1.7 times as large as Community 2.

One of the basic things these tables can teach is that mortality and fertility are closely related, and that changes in either take some time to have an effect on the total numbers of a population. They do have a more immediate impact on specific age groups of the population. Here are some points worth making:

Year	Community 1			Community 2
	Total	Male	Female	Total	Male	Female
1970	2,100	1,045	1,055	2,100	1,044	1,056
1975	2,360	1,117	1,183	2,431	1,212	1,219
1980	2,663	1,330	1,333	2,780	1,390	1,390
1985	3,024	1,513	1,512	3,127	1,567	1,560
1990	3,461	1,731	1,729	3,460	1,735	1,725
1995	3,996	1,998	1,998	3,756	1,883	1,872
2000	4,620	2,309	2,311	4,008	2,010	1,998
2005	5,338	2,668	2,669	4,256	2,134	2,122
2010	6,184	3,093	3,091	4,499	2,255	2,244
2015	7,176	3,592	3,584	4,718	2,366	2,353
2020	8,227	4,416	4,131	4,899	2,456	2,443

• Fertility and mortality are linked. High fertility also brings high mortality. High fertility implies that women bear early, often and/or late in their reproductive lives, and all three are known to contribue to high death rates among women and children. This is especially true for the poor. When we reduce fertility, especially by increasing the age at first birth (above 20) and the interval between births (more than two years), we get lower levels of infant and maternal mortality.
• Population has a momentum. The people who will give birth to the next generation, for the next 45 years, are already born. Thus even a rapid decline in fertility will not show up in reduced total population, in the case of our examples, for 25 years. Perhaps this is one of the most important lessons we can learn. By the time population-induced environmental degradation is experienced, there is still a 'built-in' population growth for the next generation.
• The impact on age groups differs. The numbers of small children and infants, ages 0 to 4, who are dependent and need nurturing and good health care, will begin to decline within five years of fertility decline. As the number of infants declines, the differences between the two communities grows very rapidly. Within just 25 years (1995 in the chart) the community without the fertility decline will have twice as many of these dependent infants and children as the community with rapid fertility decline. In 50 years they will have three times as many. For ages 15-19, the ages when children become adults, the differences do not appear for 25 years after the beginning of the fertility decline. By 35 years, however, there is a marked difference, and in 50 years, the larger community will have almost three times as many young adults for whom it will have to find education, jobs or land.

Resource: Demproj demographic projection software

One very useful program for projecting populations is called Demproj, for Demographic Projection Model. It is a full-featured population projection program developed by the Futures Group International with partial support from the US Agency for International Development. The program runs on DOS, and requires only 770K of disk space and 640K RAM. This is well within the range of most computers today, even lightweight laptops.

The strongest feature of the program is its ease of use. It requires relatively little initial input of data, gives clear choices in different types of assumptions and can produce output either in tables or in graphs. It does this in part because it is linked to a number of basic life tables, from which it computes very quickly the implications of the base population and whatever changes in assumptions (about death and birth rates, migration or AIDS) the user makes.

Requests for copies of the program or more information should be sent to:

The Futures Group International
80 Glastonbury Boulevard
Glastonbury, CT 06033
USA
Tel. +1-203-633 3501
Fax. +1-203-657 9701
E-mail: j.stover@tfgi.com

The cost is US$25.00 for manual and diskette (in 1997). The package is free to some users in developing countries.

C.2 Using a computer for map-making and GIS

Upon completion of the mapping exercise described in Chapter 4, it may be useful to transfer the paper maps on to a computer mapping or geographic information system (GIS) software package. Both computer mapping and GIS packages allow the user to create digital maps that can easily be updated, changed or manipulated. Another common feature of most mapping and GIS software packages is that point, line and area features (e.g., water courses and wetlands, farmer's fields, locations of wells, roads, houses and buildings and other infrastructure) can be stored in separate layers or coverages for easy creation and updating. The layers can then be added together to create a composite map that shows only the desired features, and is not cluttered with irrelevant information. Figure C.1 provides an example of a 'manual' GIS using acetate sheets for overlays on a standard topographic sheet. Data from such a manual overlay exercise can easily be digitized into 'layers' or 'coverages' for use in an automated mapping or GIS program.

GIS software packages perform many of the functions of a computer mapping program, but also allow the user to create a database of geographic features and to perform spatial analyses of varying degrees of complexity (depending on the sophistication of the package). The database might contain information such as the size of farmers' fields, the number of residents in each household, the literacy level of different villages or the straight line distance between points on the map. More detailed information on natural resources - such as forest cover, soil erosion data, presence of specific habitats or species or risk of flooding - can also be stored and shown. Even management data, such as the classification of forests by management regimes or the presence in a village of a management committee, can be mapped, allowing interesting overlays between socio-demographic, environmental and 'action-oriented' data.

An example of a sophisticated overlay analysis is presented in Figure C.2 (note: the map is reproduced here in one color from a full-color original; the key is therefore illegible). The map, which presents the threats to biological diversity in the Sierra Nevada de Santa Marta of Colombia, synthesizes an array of data on population distribution (including types of settlements - indigenous, campesino or large land holders - and the land uses associated with each type), roads and infrastructure, hydrology and the concentration and distribution of biodiversity. Such a map offers an immediate guide to setting priorities and scheduling action.

Spatial analysis functions such as overlay analysis (identifying the area in which there is overlap between two features), trend-surface analysis (displaying highs and lows of a particular variable as a 'draped' surface) or analyses of spatial variability (the degree to which there is clustering or dispersion of a map feature) are common features of GIS. The most sophisticated GIS packages require the computing power of a UNIX workstation and can perform hundreds of spatial and data analysis tasks. Simpler packages can run on a basic laptop computer. Thus, GIS has become a valuable tool for natural resource management as well as for a multiplicity of regional planning and location-analysis applications.

Although it may be beyond the financial means of most field-based PAR exercises, global positioning system (GPS) units can be used to produce highly accurate maps that show exact locations of specific features. GPS units send signals up to satellites that orbit the earth and, through triangulation, are able to define within 50-100 meters (depending on the unit's accuracy) the latitude and longitude of a point on the ground. This can be particularly useful in work with local groups such as indigenous peoples who need, for political reasons, to accurately define their territory (see Poole, 1995).

Computer-generated maps may be particularly useful in communicating the results of a participatory appraisal to decision-makers in government institutions, as they are clean and easy to read, and may be perceived (rightly or wrongly) as 'more valid' than hand-drawn maps.

Resource: Map Maker for Windows

Map Maker is a simple geographical information system (GIS) designed to allow non-expert users to create and manipulate maps. Using a variety of tools you can navigate around the map, measure distances and areas, draw polygons, lines and symbols, and display and edit data. Graphic objects on the map may be related to information in a data base, and data bases may be created directly from the map. You can print maps directly on to any printer or plotter fully supported by Windows 3.1 or 95, and you can produce images for inclusion in documents produced using Windows-based word processors.

The Map Maker project is designed to promote the wider use of maps as management tools, primarily in developing countries. This IBM-compatible Windows-based software is available free to non-profit-making institutions, students, and academics. The software has largely been developed in the field in collaboration with several institutions, in particular IUCN - The World Conservation Union, the Centro Agronomico Tropical del Investigacion y Ensenanza (CATIE), and the Asia Desk of the United Nations Center for Human Settlements (UNCHS). There are currently registered users in 110 countries.

For organizations or individuals in developing countries, Map Maker can be obtained free of charge. Those who have sufficient resources are encouraged to make a contribution of $50. For more information, or to obtain the Map Maker program and documentation, send a request to:

Eric Dudley
64 Tenison Road
Cambridge, CB1 2DW
UK
Fax: +44 1223 350349
E-mail: eric@mapmaker.com

For users with more advanced mapping needs there is an enhanced commercial version of the software available called Map Maker Pro 2.0.

See: Figure C.1 Manual geographic information system (planning forest management initiatives in India)
From: Poffenberger et al., 1995

See: Figure C.2 Overlay analysis produced using a sophisticated geographic information system (threatened areas by watershed in Colombia)
From: Fundación Pro-Sierra Nevada de Santa Marta, 1997