Projects & Advocates

• Financial Trends -- Smita Saxena
  • What I'd like to build:
    • I've come to realize that the reason web-based investing has not worked in the past is because there is a missing human element. I would like to create a platform that would allow real-time interaction between an investor and their advisor. Think of a basic version of Google wave where RIA's can recommend portfolios and users can either automatically execute them or engage in a 'conversation' in order to get the portfolio that they want. Once a portfolio has been executed, the user can always come back to the platform in order to look at performance and/or to make any adjustments. The RIA will also have access to this portfolio (but cannot make any transactions) and wants to keep their investor engaged enough in order to keep maintaining their portfolio through periodically rebalancing.
  • Implementation:
    • The financial roadmap tool I am proposing brings together three key services provided by the financial institutions at this point of time: creating an investment profile, getting investment advice and executing portfolios. The most lucrative of the three is executing portfolios, which is going to be the main focus of this project.
    • The plan is to link with the APIs of discount brokerage houses such as TDAmeritrade and Interactive Brokers, LLC, in order to put through trades for our investors. The trick is to have the right incentives in place in order to get users and RIAs to use our platform in order to put through the trades.
  
  
• Student-to-Student Services -- George Tang & Xuwen Cao
  • What we want to create is a Stanford Student Services web platform that connects the students with need with the students that can provide. Its a way for some students to earn a bit of money and other students to get things done quickly and on a trusted platform.
  • Feature wise there are obviously a lot of ideas to explore, some of the ones to think about are
    • review system (like the amazon seller feedback system, where you review the service that you were provided)
    • location based solutions (it might make sense to connect people closer on campus with each other)
    • partnering with places like the GSB lab / Psych lab to get more partners
  
  
• Cite Audit -- Ethan Lozano
  • Cite Audit is service that confirms and computes the validity of a claim.
  • Confirmation:
    • assists in finding support for a claim
    • gives users direct access to support for easier confirmation
    • organizes community confirmation of a claim through discussion boards and other services
  • Computation:
    • computes the scaled likelihood of a claim's validity given a variety of factors. These factors include, but are not limited to text analysis of the claim and its support, user approval of claim/support and support recognition/authority.
  
  
• Real Time People Search -- Roshan Sumbaly & Rahul Gupta
  • Most searches online on search engine tend to be people centric. Currently the trend is that people either start searching on generic search engines like Google and on finding no information switch to other sources like LinkedIn and Facebook. Also most of the information presented by these generic search engines tend to be outdated and even give links to irrelevant sites (like to a newsgroup where the person may have asked a question). We intend to build a black box which given a person's name, combines information from "profile-centric" data sources (Social Networks) to build a unified profile.
  
  
• Web based IDE -- Erik Kuefler
  • A major trend in the software industry is the migration of traditional desktop applications to web-based "cloud applications". Webmail clients like Gmail have grown in popularity compared to desktop mail clients like Outlook, Google Docs provides Microsoft Word-like document editing in the cloud, and services like Lala provide a potential replacement for iTunes. One class of application that has not seen a viable web implementation is that of code development tools such as Eclipse. The web provides many advantages for such tools: source code can be edited collaboratively in real time in addition to traditional source control, code bases can be accessed from any computer, and the ability to run server-side static analysis tools means that even a netbook could effectively run heavyweight tools like Valgrind that a high-end workstation might not be able to handle. This project proposes to begin the implementation of some basic components of such an application.
  • The client and server are both largely separate problems that will pose their own sets of challenges. On the client, we must develop a widget capable of displaying and editing source code with all the niceties one would expect, such as syntax highlighting and annotations. We would also have to develop the rest of the interface in an extensible way that would make sense inside a browser. Finally, we would have to figure out how best to represent the source code document and communicate changes in it back to the server. On the server side, we must store and retrieve documents for the client as well as providing hooks allowing the user to run external tools such as compilers or static analysis tool (e.g. FindBugs). The server should be able to run these processes in the background and send the results to the client once they have completed.
  
  
• Language Suite -- Ashwin Purohit
  • Smart Language Learning:
    We will create a GWT interface for a user to enter a word in one of many supported languages, and have it translated to the language they're learning via a freely available web translate API. To help solidify the new vocabulary in the user's mind, the returned result will include an audio pronunciation, definition, picture, and example sentence where possible. Each user's vocabulary will be stored in a list and compared to other users' vocabulary list to generate "smart" suggestions to learn new vocabulary words.
  • Native Speaker Detection
    A venture into signal processing and tying MATLAB to Java backends, we will initially define an audio corpus of pronunciations from our developers with different native languages. After drawing the relevant statistical information about the audio characteristics of different languages from processing with MATLAB, we will attempt to accept user-generated audio files and perform a nearest-match to predict the user's native language.
  
  
• Facebook FriendRank -- Shalin Mantri & Nikil Viswanathan
  • This Facebook Connect application enables people to understand:
  • Best Friends
    • Ranked list of your "best friends" as defined by various metrics
    • The logged in user can Post a Rank to a friend's Facebook profile, e.g. "Hey Max! FriendRank thinks you're my 6th-best friend. What do you think?"
  • Who's a Better Friend?
    • This feature takes as input 1 other friend for whom you want to find out: Who's a better friend? How lopsided is the friendship?
    • The logged in user can Call Out a Friend on their Facebook profile, e.g. "Hey Nikil! I've crunched the numbers on your friendship with Shalin. It seems like he's putting in 82% of the effort. What do you think?"
  • Who's Popular?
    • Ranked list of who's popular in your immediate friends' circle and in your networks
    • The logged in user can a friend's popularity rank on their Facebook profile, e.g. "Hey Jan! You're the 4th most popular person in Natasha's network. How popular is Natasha in yours?"