The first challenge is this project was acquiring the data. The ESRI street data chosen for the project was brand new and supposidly "became available" midway through the semester. It did not actually arrive until almost a month later. This is a somewhat unusual problem due to the "cutting edge" nature of the data. However, if faced with a similar situation:
The data is voluminous-- not surprising considering the level of detail it contains. The east texas street data alone is 182MB. (!) It was fairly easy to get around this problem in this project by clipping the data to the small area being studied. However, if dealing with larger area, it may be necessary to select only the larger roads, although this would limit geocoding ability. Additionally, to make ftping practical, larger street files had to be paired down for the exercise.
Some addresses are difficult to geocode. Several "No Matches" resulted from the original shot at geocoding the addresses. Some of these can be accounted for when performing an "Interactive Rematch" and resetting the Preferences...
Some addresses matched with a different spelling, zip code, or street preface (e.g. "N" vs. "S") Without being familiar with the area, it is difficult to ascertain if these "corrected" addresses are actually right. It could be that the street database does not contain the road where the industry is located, or it is "UNNAMED" in the database. (Many are)
Most of the No-Matches were the result of P.O. Boxes, which obviously can not be geocoded. In these cases, actual street addresses have to be acquired from other sources. This can be accomplished in some instances by using Yellowpages or calling the company.
Some addresses flat out do not geocode. In this case, the closest cross streets could be located and used as the point location. Depending on the accuracy needed for the application, this approach may or may not be appropriate. Additionally, the SPLAT datbase contains latitude and longitude information for some industries which do not geocode. These points can be added as a theme as instructed in the exercise. In any event, especially in this application, it is important to ultimately perform quality control and confirm questionable addresses by other means (e.g. checking against a paper map, calling the company etc) to ensure accuracy.