Brief Note on Mashups
I hope all of you in the United States had a nice Thanksgiving Holiday and all of you outside of the United States had a nice weekend. I took a long (for me) break from blogging and work to celebrate and be with my family. This year we did not make the 10 hour Griswald-ish trip to St. Louis, we stayed in Cleveland and feasted with my inlaws. I brought the Christmas lights down from the attic, they are still in their boxes by the front door.
I recently finished and submitted my article to MRSQ on mashups. Here is a brief synopsis about the article and what I learned writing it.
Mashups are blended applications combining two or more existing programs to produce a third program. To simplify, think of it as the technological equivalent of making brass. You have two separate fully functioning programs that have specific properties and do specific job, such as the case with copper and zinc. When you combine copper and zinc you get brass. When you combine two (or more) computer applications to for another program you get a mashup.
As is the case with combining copper and zinc to create brass, your supplying data sources are key to your end result. If you combine 60% copper and 40% tin you actually get bronze not brass which is similar but different than brass. If there are any impurities in your metals you will either have to select a new source metal or work very hard in the firing and hammering the impurities out. This is the case with mashups. Your resulting program is only as good as the programs supplying the data. If you have a flawed data you will have a flawed mashup. If you have hard to extract supplying data, your mashup may not be as flexible and as timely as you want it to be.
Like all new technologies, the rules and standards have not quite kept up. There are ongoing issues and questions with data such as security, privacy, copyright, and licensing. If anybody is curious about the security and privacy of data in mashups I highly recommend reading "Data Mining 101: Finding Subversives with Amazon Wishlists" where Tom Owad was able to create a list and view a map of their home of people who had "controversial" books on their Amazon.com wishlist, all from publicly available data, no extra money, and 30 hours of his time.
In addition to security and privacy other issues regarding the availability of necessary data is a central theme for some mashups. As Declan Butler (creator of Avian Flu mashup) mentions in his blog, not all data is easily available or in machine readable format. Thus certain mashups might be difficult to create or maintain based on the data that supplies it. Think of all the biomedical and scientific databases we libraries have to pay to subscribe to and all of the potential data within those databases.
However, this is just the beginning for mashups. With the emergence of the Talis Mashup Competition, OCLC's Research Software Competition, and Mashup Camp, expect to see many more mashups coming in the future in mainstream life as well as in medical and library arena. Perhaps you are already using a mashup and don't know it. Take a look below to see some of the mashups I stumbled upon while writing the article.
Everyday life mashups:
- 1001 Secret Fishing Holes -(probably not so secret now) Over a thousand fishing spots in national parks, wildlife refuges, lakes, campgrounds, historic trails etc. This mashup uses data from Google Maps and Recreation.gov.
- SecretPrices.com -Comparison shopping site with deal, rebate and coupon information. Customer reviews and product information from Amazon.com and Epinions.com. This mashup uses data from Amazon.com, Shopping.com, and Epinions.com.
- Zillow.com -An online real estate service providing you tools and information on home prices, locations, etc. This mashup uses Google Maps and provides APIs of its own data for others to use.
Medical and Scientific Mashups:
- Avian Flu Mashup -Maps the spread of the avian flu worldwide including human cases and poultry outbreaks. The map also provide additional data on each event, and additional datasets, such as poultry densities worldwide. *You must install the latest version of Google Earth (for complete instructions go to http://declanbutler.info/blog/?p=58) This mashup uses data from Google Earth, WHO, and UN Food and Agricultural Organization.
- Vimo -Formerly known as Healthia, allows users to search for information on doctors, hospital costs, and insurance plans. This mashup uses "variety of private and public data sources so that shoppers can find a physician and compare hospital prices for medical procedures."
- PubWindows -Searches PubMed as well as three other PubMed text mining tools: Chilibot, XplorMed, and BioIE, while also including SFX linking information for two institutions.
- Umlaut -Winner of OCLC's Research Software Competition, very cool open source mashup (so you can grab it tweak it and use it at your library) for finding full text articles, books, and information. It combines information from an OpenURL Link Resolver program, with web content from Amazon.com and WorldCat.org.
- Book Burro -Web 2.0 extension for Firefox and Flock. When it senses your are looking at a page that contains a book, it will overlay a small panel which when opened lists prices at online bookstores such as Amazon, Buy, Half and whether the book is available at your library.
- Go-go Google Gadget -Using Google Gadgets API it allows patrons to put customizable panels on Google's personalized home page containing library OPAC information such as books checked out, new arrivals, hot items, and requested material.