This edition of the Relation Browser (RB) enables searches of 10,900 assistive technology products. The product information is collected from ABLE data, which maintains an online database of assistive technology products. The data in this RB is a snap shot (October 2003) of the ABLE database.

My snap-shot of the data might not reflect an accurate view of the ABLE data. The database is continuously being updated. My collecting and processing of the data might introduce errors into the product information.

Requirements: A browser running JAVA 1.4.1. If it is installed, the RB should appear in a separate browser window. 

How to search the RB for assistive technology products:

There are two methods to specify a search, which can be used together or individually:

1. The user narrows the product search by selecting description in the categories that are apply. 
2. In the middle bar the user enters a search text in the appropriate fields.

After the search is specified press the “Search” button to retrieve results. The results appear in the lower panel. Clicking a search result automatically loads the ABLE data web page of a search from the clicked product. Click the product name on the ABLE data web page to obtain more detailed information about it. 

Press the “Restart search” button to clear the search results. 

The interface enables searches of assistive technology products. It consists of three panels
 

• Top panel is an overview of the products by category 
• Middle panel is narrow and lets the user enter search criteria and a button to start the search.
• Lower panel lists the search results. 

The overview categories in the top panel are as follows:

• Disability Category: Type of disability, activity or place the product assists a user. Products are listed under different categories. 
• Price: Price of product. Some are unknown, or range from free to over $1000.
• Year of price quote: The year the price quote is obtained by ABLE data. 
• Region of Manufacturer: The states in the regions are as follows:

Click image to enlarge:

The fields in the product search are:

• Product name
• Detailed description
• Price
• Manufacture’s address in detail
• Comment 
• URL to ABLE DATA information on product. The link is actually a search of the ABLE Data database, so the user is required to press the product name on the ABLE Data web page. 

Trends recognized in the overview panel:

Lessons learned about assistive technology products:

• Usually assistive technology products are considered expensive but the overview indicates many products are below $200.
• Several products are “do it yourself”.
• Manufactures of assistive technology are distributed through out the country. However, comparing absolute numbers is not possible because the regions represent different size geographical and populations.

Information ABLE data developers:

• The majority of price quotes are from 1999 and after 5 years it might be useful to update the prices to reflect changes. 
• With the overview it is easy to search for products with incomplete price or manufacture information. Those products are found by searching for products with price “unavailable” and manufacturer unknown. 

My experience is that it is easy to work with the RB. More documentation would be helpful. Here is what I learned:

• Attribute values should be 1 or greater.
• Attribute ID and dataID should be ints
• To load in strings longer than 255 characters use the BLOB field in MySQL
• My html file demonstrates how to set user-login and password when the two are not identical. If the password field is missing. 
• The year data is in unusual order 2000-2004, before 1995-1999. The years are sorted in order of attributeID and not orderBy. 

I used a python program to crawl the ABLE data web site. It crawled the following web pages and terminated when the detailed product information is recorded.

The pages are as follows:

1. Search by product type
2. Category page, example Blind and Low Vision
3. Sub category
4. List of products (click link from previous page)
5. Single product page (click link from previous page)


It took 20 hours to record my data set of 100 MB. Product entries, however, were duplicated. 

The collected data was saved to text files. I wrote and ran more python programs and Excel macros to remove duplicate, format fields, calculate attribute values for the RB categories {category, price, year of price quote, region of manufacturer}. 

The final text files are uploaded into MySQL using the following command for each table:

 
LOAD DATA
LOCAL INFILE <file name of tab delimited file>
INTO TABLE <MySQL table name>

Example:

LOAD DATA
LOCAL INFILE 'D:\\Python22\\code\\inls357\\int1up_upload\\att_year.txt'
INTO TABLE 1_ATT
 

The RB URLs are links to a search of the ABLE Data web-site. Individual products on the ABLE data web-site do not have a fixed link but are dynamically generated by the database. The dynamic link has an expiration and therefore cannot be used in the RB.
 
 



Note: The Relation Browser should appear in separate window