Input:
The Amazon Book Explorer uses the Product Advertising API to retrieve a list of 100 books from Amazon.com's database and uses the data from these titles to provide different analytics.
Enter a query to search for books.
Select a ranking method for your query (if there are more than 100 results, e.g. "sales" will use the 100 best selling titles, "customer review" the ones with the best customer reviews, etc.).
Select a local Amazon page.
Select the number of pages (one page is 10 results, max. 10 pages can be retrieved at the moment).
Output:
The tool currently provides two analytical methods, both putting out graph files in the .gexf format, for use with the
gephi graph visualization toolkit.
1) Co-word anaylsis of book titles
If two words appear in the same book title, they are considered as “linked” and the more often they do, the stronger the connection. Trough network analysis, patterns can be made visible.
2) Classification mapping
Every book in Amazon's database is placed in one or several subject categories (e.g. hobbies => gardening => flowers). The tool collects all of the category trees and attributes a frequency to every node. A subject – via the query – becomes localized in a subject organization system and, through the frequency count, the weight of individual categories can be made immediately visible, e.g. by sizing nodes according to frequency.
This sample project shows the two types of output for the query "retirement". First, the co-word analysis, with most of the less significant words filtered out. Node size is "degree" and node color "betweenness centrality":
Second, the mapping of category trees and frequencies. Size represents the frequency of a category, color the level of placement in the hierarchy of categories:
![amazon_categories_retirement.png amazon_categories_retirement.png](https://wiki.digitalmethods.net/pub/Dmi/ToolAmazonSampleProject/amazon_categories_retirement.png)