Lab Report #2a: Google Fusion Tables
I am attempting a map visualization of quilts created in the first half of the 19th century. I am attempting to code the quilt patterns on the map with different icons, to enhance cognition so users can discern patterns and make observations.
Dataset exported from The Quilt Index – I searched for occurrences of the 22 patterns Gracie Mitchell used during the years 1800-1849. The Quilt Index, a free open-access project of Matrix, Michigan State University Museum and the Quilt Alliance, is an online database of metadata about quilts in the collections of several different museums, universities, and research institutes throughout the country.
MS Excel, Google Refine, Google Sheets, and Google Fusion Tables
Gracie Mitchell’s interview transcript notes, recorded by FWP WPA interviewer Bernice Bowden in 1938
- Searched the Quilt Index for each pattern, saving items to compare, viewed all of the records presented in a table for comparison, then cut and pasted (transposed) the fields into Excel.
- Used Google Refine to clean up all of the variations of names used for the patterns, the quilts (items), and the locations.
- Uploaded the file in to Google Sheets.
- Opened Google Fusion Tables and imported the Google Sheet in to that program, which gave me a pretty good map to start with, but all of the icons are the same color – red.
- ‘Googled’ how to create separate icons using Fusion Tables and learned that I needed to make a new Google Sheet which defined a specific icon for each pattern type. I tried to pick icons that related to the name of each quilt pattern.
- Fused the two sheets creating one “merged table”, which has the ability to filter, and have a map that uses a different icon for each quilt:
I was pleased with the map that I was able to make with Fusion Tables, because I have been struggling to use Exhibit for over a year. The map shows all of the patterns with a different icon, and the user can zoom in and filter. As we discussed in class, when the location given is only the state, the program piles a bunch of item icons on top of each other. There needs to be a way to tweak the size of the icon. Also, regarding icons, I chose images that somewhat related to the name of the pattern (Drunkard’s Path is the martini glass), but the best imagery for this map would be the quilt block. Also, this map is very cluttered, which means that the dataset for 1850-1875 (741 records right now) will be overwhelming. In the end I will definitely have to create small multiples in order to facilitate analysis.
Now that I have a better understanding of the classification system for quilt patterns, I think that I will group some of the patterns so that I can use 5-6 colors of dots on the large map. Then I will probably create small multiples using screenshots of ideal portions of the map for each pattern. I will also work on creating small quilt icons or thumbnails based on my research using the Brackman Encyclopedia for the timeline.