TechnologyIn the area of technology, we focused on:
|
|||||
VisualizationWe experimented with several different ways of representing data on a two-dimensional map, from a simple choropleth map to micromaps linked to box plots and spatial box plots. In addition, we also experimented with 3-D technology. Samples of these products are provided below; links to more information regarding this work are also provided. |
|||||
|
|||||
Software Tools
We developed several software tools to assist us (and others) in the process of data analysis:
Since most of the data we collected and analyzed in this project originated from large databases, we received this data in a "relational table" format. As with any data structure, the relational format has some advantages (efficient and reliable for data storage and retrieval) and disadvantages (format is generally not compatible with standard spreadsheet or statistical software such as Excel, SAS, and others). Relational Table Format (Sample)
Note: 1) one observation per row; 2) difficult to summarize data (e.g. all nitrate values) without rearranging the data In order to analyze data in a relational table format, one must usually rearrange or transform the data into "traditional" spreadsheet format. Traditional Spreadsheet Format (Sample)
Note: 1) one date per row; 2) data can be easily summarized by column There are several options for transforming data, including:
All of these methods have some drawbacks: manual transformation of datasets (especially large datasets) is tedious, time consuming, and error-prone. Pivot table and cross-tab query methods are much more efficient and accurate; however, only one row and one column heading can be selected to appear in the transformed table . This becomes important in the context of row headings (or indices) when more than one field is necessary to uniquely identify a sample (for example, the date and location of a sample). The tables above were transformed using the Excel Pivot Table; note that the station names have been inserted to the left of the tables. In the context of column headings, especially when a table contains data measured in different units, it is necessary to know which units (e.g. milligrams/Liter, micrograms/Liter, picocuries/Liter, etc.) correspond with which contaminant(s)--in addition to the name and/or abbreviated name of the contaminant. Note that the units are absent from the transformed tables--either the contaminant name or the units (but not both) could be selected with the Pivot Table or Cross-Tab Query. The first software tool we developed, therefore, was a data transformation tool, which allows the user to specify (based on the original dataset) what row and column label(s) he/she would like, and output the file into a text file. This file could then easily be imported into Excel or other spreadsheet/statistical software of choice without loss of vital information (i.e. units) Transformation Tool Menu
|
|||||
Regression Tools |
