App Ideas
From Semantifi Wiki
We are building a few initial Apps to demonstrate how other publishers can build many more Apps on SEMANTIFI platform.
There are an estimated 115 million datasets on the Web. No one company can Semantifi them all. That's where you, The Publisher, come into the picture. If you have datasets that you would like to investigate then configure them to SEMANTIFI and publish them as Apps. Currently all Apps will be available publicly, for everyone to search. We are working to allow you to share Apps within groups and, if you choose, for a fee.
Here are a few ideas for building Apps that we certainly will be excited to see on SEMANTIFI. These ideas are generated by our team and hope you and the community can share your ideas as well. Lastly, on building Search Apps, You can do it and we can help (like they say at the Home Depot).
Contents |
App to Search Wikipedia
Challenge: Wikipedia currently only supports keyword-based search and does not allow more expressive queries like “All cities in New Jersey with more than 10,000 inhabitants” or “Italian musicians from the 18th century.” This lowers the overall utility of Wikipedia.
Value: A SEMANTIFI App can allow Wikipedia users to ask simple questions and get relevant knowledge base driven results. This can revolutionize access to this valuable knowledge resource.
Solution: This can be achieved by configuring community built knowledge bases like DBPedia to SEMANTIFI. Wikipedia covers a broad cross section of human knowledge base and often referred to as the “nucleus of the webs knowledge”. Before DBPedia, building an ontological framework representing all of Wikipedia has been deemed impossible or very difficult for any single company or group of experts to build. DBPedia got around this challenge by systematically extracting the details inside infoboxes from Wikipedia pages to build a broad ontological model allows knowledge base driven search platforms like SEMANTIFI to deliver semantic search on Wikipedia content possible.
Open Linked Data Apps
Like Freebase and DBPedia Apps discussed above, there are dozens of linked datasets that are waiting to be searched with a Semantic Search Engine. Today, these datasets are either searched by keywords or through SPARQL queries. If you are wondering what SPARQL is, don't worry we won't be forced to learn it, it is one of these programming languages for querying data triples or RDF databases or simply graphs. We particularly like Crunchbase and hope to build an App soon. Check openlinkeddata.org and see what interests you.
Government Data Apps
Vivek Kundra, CIO US Gov, followed up Barack Obama's campaign promise to launch DATA.GOV with the goal to make government data transparent. Data.gov has nearly 10,000 datasets and it is expected to have a million datasets very soon. To make data transparent, "raw data" must first be freely accessible and second it must be searchable. Data catalogs like Data.gov are taking aim at the first. Semantifi helps anyone to make them searchable.
US Census App Who would like to investigate population and demographics trends? It only affects our everyday lives -- especially if you are a consumer products company looking for your niche or an advertising company helping your clients understand these trends.
Government Spending App FedSpending.org had over 10 million searches in less than a year of launch. Who would have thunk that we care about our government beyond the once-in-4-years-presidential-political-ritual?
Check our Govt Spending App on SEMANTIFI which makes it very easy to search where and how our taxpayer monies are spent.
Congressional Earmarks App We should be able to research what sort of pork our representatives and senators attach their names to. It doesn't have always to be a factoid on CNN, Fox News. We can research, discuss and engage in democracy if the data is accessible (kudos to TaxPayers for Commonsence for that) and searchable (you can build an App).
Dashboard of Obama App How about a Dashboard for Obama? It can show Fed Spending data, Tarp/Bailout data, Recovery data, and allow citizens track the progress of huge government initiatives.
Apps to Share Global Research Data
Challenge: Take any specific cancer research community and there may be hundreds or even thousands of researchers working around the world across academia, non-profit and for-profit organizations but progress is terribly slow. The typical process of collecting data, analyzing it, publishing research and sharing with the community is MULTI-YEAR process. I understand this firsthand as a co-author and analyst for State Dependency Learning and Drug Discrimination studies published by Dr. Donald Overton. I helped analyze these datasets some over 15 years ago in my graduate program and only last year were they published.
Value: Any researcher can publisher their datasets as Semantifi Apps and share them privately within the research community or openly with public to begin the research community discourse IMMEDIATELY not after a decade or two.
Imagine the global research communities across energy, environment, disease, treatments, global warming, etc. Just imagine, the transparency issues with global warming research that have come up recently could have been addressed by making the data widely accessible.
Scope: Government and research data is estimated to account for nearly 50% of the Deep Web. These datasets may not offer commercial value to its publishers but they can create immense value for everyone in the research community.
Craigslist Apps
Challenge: Craigslist currently only supports keyword-based search queries and the search experience can be significantly improved with semantic search.
Value: Multiple Apps are possible across the different verticals like Jobs, Autos, and so on. With Craigslist Jobs content, a recruiter can ask “Java programmers with over 5 years experience living within 60 miles of NYC” and get meaning based results and save hours in screening compared to keywords matched results.
Using Craigslist’s Auto Classifieds, a buyer or seller can ask questions like “Honda Accords with less than 10,000 miles and under 3 years of age” and save hours with relevant results in finding the deal. Craigslist allows third parties to crawl and index its content. Armed with vertical specific knowledge bases, Third parties can apply SEMANTIFI to deliver far more relevant results compared to Craigslist’s own keyword search results.
Solution: Since Craigslist's content is not in a database, additional tasks are needed to deliver semantic search on it. It includes building a knowledge base (moderate amount of work), crawling the content (simple), and building a semantic index of content (we can help), and then apply SEMANTIFI.
Amazon Product Search App
Challenge: While Amazon offers the best shopping experience of any retail web store there are opportunities to improve this shopping experience further through better search experience. Currently, search at Amazom.com is keyword-based and is complimented by guided summarization “left-hand-groupings” but users cannot ask simple questions like “laptops under $450 with XP Pro". Today, such a search query produces "zero results or not very relevant results". It requires several actions like selecting many categories, defining price ranges and so on.
Value: A SEMANTIFI App on Amazon Product Catalog can present highly relevant results to such questions and significantly improve the overall shopping experience. The value from such improved shopping experience is quite obvious.
Solution: Amazon provides access to its entire product catalog via Amazon Web Services. Amazon not only encourages third parties to build custom stores powered by its product catalog but also fulfills the orders using Amazon's own payment and order processing infrastructure to ensure customers confidence. Lastly, Amazon offers upto 15% in sales commission giving the incentive to build custom stores.
There are many examples of stores powered by Amazon product catalog. http://www.zoomi.com is a good example where an innovative presentation approach is all that is needed to profit from a custom store. Similarly, a store with semantic product search powered by a SEMANTIFI App can be compelling to a large enough number of shoppers to build a profitable custom store. Lastly, Amazon itself may want to apply this search capability on its own store.
Amazon Marketplace Sellers App
Challenge: Amazon’s unrivaled product breadth comes from the many sellers peddling numerous, often less common, products through Amazon Marketplace. However, these sellers do not have seller tools built on historical sales data like eBay Sellers do via tools from Vendio, Terapeak, etc.
Value: Seller tools can help understand what to sell, when to sell, where to sell and how much to price. Sellers can research most successful products in any category, learn sales levels, average sales prices, and the highest possible prices, understand seasonality, know what categories will get the most attention and so on. This presents a win-win opportunity for Sellers, Amazon.com and you the App Publisher. With access to Amazon's historical sales data, you can build and support Seller Tools via a SEMANTIFI App. Amazon can promote it to Sellers. Sellers can subscribe to these tools at Amazon.com itself.
Movie Search App
With a App searching Movie databases like IMDB.COM, etc. users can ask such questions like “Movies Directed by Clint Eastwood” and get only those few movies he directed instead of all movies with a mention of Clint Eastwood. As for benefits, you can build the App and offer it as a service to search portals, movie portals, and numerous websites instead of licensing movie content and transferring data.
Social Network App
If you are really ambitious, build a social network. While Facebook is the new kid on the block, Myspace used to be just that a couple years ago and Friendster before that. This was the sentence written about a year ago. Now, it's Twitter which is the new kid on the block. None of these networks offer a "lasting" value proposition like Google has brought to the search space in spite of not being the first(is that Yahoo, Ask, Excite, dime-a-dozen, etc.) or the newest kid (that title goes to Bing!).
An exciting App built using SEMANTIFI is a “search engine built social network”. The idea is to crawl the web and intelligently connect people based on their relationships, not just contacts. Vast amounts of personal information either self-published or published by others is available online. Crawling the web offers rich information about people’s relationships, behaviors and interests which simply cannot be obtained from contact lists. Then users can ask questions like “Colleagues of Carol Bartz at AutoDesk” to get relevant results. While in the past this project would have required millions in funding, now it can be completed at a fraction of that cost given search services like Yahoo Boss, Amazon Web Services and semantic search platform to build Apps for free.
