ANALYZING METHODOLOGICAL & OPERATIONAL ERRORS       

Upon doing this project, there were several errors throughout the process of my project.

Lack of Experience:

This is the biggest problem when I was doing spatial analysis.  As it was my first time doing such project, my knowledge on GIS and experiences on managing data are definitely not sufficient.  Especially when doing spatial analysis, I have to try using different methods in order to figure out the correct method to do the analysis.  I am able to find an appropriate method at last.  However, under this circumstances, the accuracy of the result would not be as accurate as if it was analyzed by people with GIS experiences.

Data Confusion:

As I have spent lots of time attempt to find a new data set, which made me realize that it is really difficult to do so.  Data is not just an image, we can see from the metadata that it is included type of projection, datum, longitude and latitude and so on. 

The problems I encountered were first of all, all data were not in raster format and the translations from vector to raster were usually failed.  Secondly, even though I can translate them to raster layer, due to projection different or row and column different which is impossible to overlay with other layers.  Consequently, all the data I am able use is only those under SIS drive.

There are plenty of data under SIS drive, but I am not familiar with all of them.  Since the name of the data is not clearly shown, I have to open most of them to find out what those data are all about.  This is time consuming and sometimes I still do not understand because the metadata are not viewable. 

Operational Errors:

-   Negative number shown in the legend after convert vector to raster, this is different than I usually do during the lab.  I have to reclass the ranges accordingly in order to obtain an accurate image.

-   In the tutorial of Idrisi 32, it mentions that distancebool and bufferbool should be identical and either can be used.  When I was operating distance and buffer, I always have two different images.  I have to try different methods, such as: reclass and reclass again to obtain the image it should be.

-   After obtaining the MCEBOOL image, all the suitable campsites are very small and unclear.  I have to change the color from palette several times to find a more readable image.  

-    The accuracy of consistency ratio in pairwise comparison file is not high when doing weighting factors for aggregation because they are judge by personal generalization.  It has to be manipulated several time in order to get a ratio less than 0.10

This is the end of my project.  Thank You for your visit.

Back to HOME