Problems:

1. Outdated Census Data:The U.S. census Data used for this particular project dates back in the year 2000, hence there is a significant time gap in relation to the time of this project was initiated back in mid October of 2006. Because of such a wide time lag, it is very likely that the U.S. Census data does not accurate describe the present population in the city of Chicago. For instance, since 2000, many things would have change as the city's Chinese population may have grown significantly due to the inflow immigration from past 6 years. Another instance may involve the case where the average number of family members in general may decrease slightly as many indivduals may decide to have less children, thus overall number of families that have more 4 or more members may be lowered. Whatever the changes maybe, no one would know for sure until the next U.S. Census is taken, which will be in 2010.

2. Population Density should be incoporated: In general, the numerical values shown when pointing and clicking on any part of particular raster images in the cursor inquiry mode in Idrisi, represent the total of number of individual that fall in the described category. For example, in the category of Family Income that ranges from $100K to $150K, various Census Track areas have various number of individuals that fall in that particular category. However the problems relates to the fact that not all the Census track area have the same dimension, as some are bigger or smaller than others. As the result even when Census track area A may have the exact or similar number of representatives of a particular catetegory as Census track area B, the MCE would favour the one that have the larger dimension. To resolve this particular problem, population density and not population should be used to represent a particular Census Track area.

3. Important Census track area(s) may be omitted when row & column values were changed: During the image editing procedure involving changing of the row & column values of individual raster images so that they can be overlayed, some census track areas were omitted. Consequently, this was due the case where the all the row & column values for raster images was reduced significantly which then decreases the overall study parameters (Width x Length). Compare the images below for a visual comparison. Such case scenario would negatively affect the accuracy of the MCE as omitted Census track areas may have contain a significant number of representative of a pariticular categories.

Raster Image with original row & column values	Raster Image with reduced row & column values

4. Reclassed values were subjectivly set: When reclassing raster images from various categories, different percentages of the total population (25% / 50%) were used to highlight Census track areas that have relatively high number of representatives. The problem relate to the case where the percentages that was used in the reclass was not based on any statistical studies but rather on personal decision.

5. Chosen factors were based on personal experience and opinion: All the factors chosen for this particular project were reflected from my personal experiences obtained from years of dinning in numerous fine Chinese restaurants located throughout the GVRD (Greater Vancouver Regional District). Whenever I dine in a particular Chinese restaurant, I would often observe my surrounding environment and see what types of people are there in general. For instance, from my personal observation, I would often see that many of the dinning patrons were relatively high income earners as they would often drive expensive imported vehicles and would often talk about their newly aquire real estate property to their extensive family members / relatives. In general, there are various potential flaws to my choice of factors. Firstly, they are based on personal experience and observation and not any extensive quantatative research study. Secondly, the observation was based on the GVRD area and not the city of Chicago. In contrary, perhaps people that dine in the fine dinning Chinese restaurant in Chicago may not be upper income earners or possess real estate properties. Until I personally go to Chicago and conduct an observational study, I will not know how signifcant and applicable are the factors that I have chosen.

6. Lost of image details during Rasterization: When the vector images were rasterized, a signifcant lost of image details were resulted. Although this does not directly affect the overall MCE, it does however affect the precision when narrowing down and specifiying the cross streets that border the highlighted area shown in the Result page.

7. More Constraints could be incoporated: Although more constraint types could be incorporated into the final base map, only four types of contraints were used. This is because the omitted constraints did not permit any changes made on their row and column values which subsequently affect the overlay process.

8. Unit Measurement error: While this does not effect the overall MCE, the unit of measurement used in the original Vector images is in feet while measurement unit used in the Raster image is in metres. Although I would like to make the neccessary unit changes in the early begining of the project, in Idrisi, such option was not permitted during the time where the projection was specified.

Previous Page Next Page

Back to Home Page