SEEN
economic
2S in rural
lity of the
rage and
pplication
| which is
ity of the
nain data
obtained
oth have
e of main
ementary
plains the
re shown
(S.
d Social
vernment
vities and
its policy
elopment
1e Prime
nal Rural
. conduct
1e socio-
as of the
-very two
' Surveys
)-district),
age, and
ut into a
titute for
Jniversity
| this was
1990, the
the 1986
jatabase,
lage-level
structure,
e and sO
and the
he topics
are given in an earlier paper (Nagata, 1996).
The administrative village, which is the base unit of the
survey, is not always coterminous with the natural village.
In Northeast Thailand, however, villages are usually of
clustered form, and the administrative and natural
villages mostly coincide. The average size of an
administrative village is 100 to 200 households and 500
to 1000 people.
22 Reliability
There are some drawbacks in the village database.
Careless mistakes such as wrong entries and wrong
estimations are commonly associated with any
questionnaire. Inevitable errors also occur when the
data input into the computer, since the survey books are
not machine-readable but hand-written. Questions
essentially difficult to answer correctly also exist.
Some careless mistakes can be checked by the internal
consistency of the data. Not only the total village
population, but also its breakdown by sex and age are
given. These allow cross-checking and, in some cases,
correction of figures. The percentage of apparent
mistakes in demographic data in the 1992 survey was
4.8%, and 90% of the mistakes were corrected.
One example of the questions essentially difficult to
answer correctly is the average annual sales of a
household which is engaged in home industry. The
related question of the number of households engaged in
this work is easier to answer, but the scale of working
varies from one household to another. It is apparent
that the surveyors do not obtain an accurate value. Yet,
they are often able to present reasonable estimates value
which can reasonably compared with those of
surrounding villages.
Thus, we must keep in mind that there are defects.
However, these defects are compensated by the
coverage of a great number of villages and the provision
of a wide variety of otherwise unacquirable information.
In one case study, the village database has been used in
an attempt to classify villages in Yasothon Province, one
of the nineteen provinces in Northeast Thailand (Kono &
Nagata, 1992).
3. NETVIS
In the light of my experience in the case study in
Yasothon Province, | have been developing a framework
with which to utilize the information in the village
database for the whole of Northeast Thailand. Many
Preparative works were required to construct the
Northeast Thailand Village Information System (NETVIS)
as a GIS application, because a digitized data set of
national geographic coordinates information for Thailand
Snot yet available. Fortunately, since the Reforestation
and Extension Project in the Northeast of Thailand of the
Japan International Cooperation Agency (JICA-REX)
Sought to utilize the village database for its target setting
and evaluation, | was able to conduct these preparative
Works as part of the project.
3.1 Components
NETVIS is composed of a unit of databases and a unit of
mapping. The former includes the following databases:
(a) single-year village databases; (b) a village position
database, which is indispensable for mapping and is not
provided with the village database; and (c) a village
identification database, which correlates the data sets of
a village in different survey years, since the identification
numbers used to tag villages are not necessarily fixed
from one survey to the next. More details about these
databases appear in an earlier paper (Nagata, 1996).
The unit of mapping includes programs which calculate
and modify the data in the unit of databases in order to
map them. NETVIS is constructed on MS-Access 2.0
on MS-Windows 3.1.
3.2 Data Mapping
The handling unit in the NETVIS is an administrative
village, as in the village database. Over 26,000 rural
villages in the Northeast are included in the village-level
survey and it is certain that errors exist. So these
conditions must be considered in mapping data.
The data maps presented in this paper were composed
by calculating an average value or an accumulated value
for each mesh. The meshes are of three-minute
intervals of both longitude and latitude, that is 5.3 to 5.4
km from east to west and 5.5 km from north to south.
The data set of village position is vectored as point data,
but the outputs are presented by raster graphics. Each
mesh contains O to 23 villages, with an average of 4 to 5.
To calculate an average in a mesh is also effective to
minimize errors. Although it is possible to vary the size
of the mesh to meet specific purposes, | have found the
three-minutes interval to be most suitable as a result of
several experiments.
4. CHANGES IN NORTHEAST THAILAND
Northeast Thailand is bounded by the Mae Khong
(Mekong) River and Laos to the east and north,
Cambodia to the south, and North and Central Thailand
to the west. To observe changes in rural areas of
Northeast Thailand in the latter half of the 1980s, some
outputs from the NETVIS are introduced below. The
locations of 85% of over 26,000 villages have been
identified and the data on them are used. Small circles
on maps below show the locations of provincial capitals.
4.1 Infrastructure
Improvement of infrastructure is one area in which
governmental efforts have brought rapid progress.
Figure 1 shows the percentage of households supplied
with electric power. In 1986, electric power supply was
available where population density was relatively high,
especially around major local cities. In 1992, only a few
areas were left unsupplied.
Figure 2 shows the percentage of households supplied
with water by pipeline. In 1986, water supply by pipeline
was scarcely available, but in 1992 an improvement can
517
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B7. Vienna 1996
