Here we go with the final 'Data Quality' focused week, wrapping up with an assessment that determined the difference per grid polygon between two road networks in Jackson County, Oregon:
The big negative differences implied that the Street Centerlines data attributes were less complete than the TIGER Road Project data attributes. The positive values resulting from calculating the difference implied that the Street Centerlines were more complete in those areas. For this analysis, being within -10% to 10% was determined to be a reasonable or acceptable difference, although overall the TIGER Road Project data proved to be more complete. Although there appears to be two areas with no value for difference in two grid polygons, one of the areas has better data from the TIGER Project versus the Centerlines, which did not display well due to a lack of a 'difference value' in the final map results.
This lab's assignment has proven to be a valuable and challenging learning opportunity. A rare experience: I ended up calculating the data successfully without being able to understand and interpret the result enough to create the visual interpretation. I started over to gain a better understanding of the analysis, see how well my results matched up, and experiment with alternate tools. It took me four tries to feel confident enough with the visual and numerical results to submit the assignment.
The importance of knowing the quality of data and showing the quality quantitively cannot be overstated, and this was a nice start in training myself to think more statistically about ensuring data quality. More to come!
Comments
Post a Comment