CS190 2006 Project Ideas



Bookworm

I'm a Brown employee and I volunteer in a public grade school. The school library is 100% supported and run by pto volunteers. A parent volunteer wrote a Microsoft Access application that uses a CueCat (the Providence Journal gave them away for free several years ago!) to scan the barcoded books and student lists. It allows us to check books in and out of the library, look up books in the database by title, add and delete books from the database (when the pto buys new books or a book is lost or mangled), print new barcodes, check to see what books a student has checked out, etc, etc.

But there are many things that need improving and the original source code has been lost, so we need to start over. We have something like 10,000 barcoded books in the library and all that data is in an Access table with title, author, category, etc. We know exactly what we want to be able to do, and would have no trouble defining our requirements. The pto could probably pay for a new barcode scanner (since the cuecat gives us trouble sometimes) and might be able to budget some funds if there was anything else that needed to be purchased.

Submitted by Cynthia Davis (Brown Employee)



Experiment Design

The goal of this project is to create a tool for running cognitive science/psychology experiments on the web. The idea is to create a user-friendly interface that would allow experimenters who know little about the web to specify the design and materials for an experiment. The system would then create a set of web pages that run the experiment, and that would save participants' responses in an easy-to-understand way that could be freely accessed by the experimenters.

The Scholarly Technology Group created such a system for my lab several years ago and we use it frequently. However, it has some critical limitations that I would hope a new system would overcome:

1. It is limited in the variety of experimental designs that can be implemented (the number of independent variables and whether they are tested within- or between-participants).

2. It only allows us to present text to participants, other kinds of images are not supported.

3. Everything appears on a single web page that participants can scroll up or down. We'd like to have more control over how materials are presented.

4. We don't have direct control over the consent form that's presented.

5. Security could probably be enhanced using more modern methods.

The code for the earlier system is available. However, I suspect that a new system would have to be built pretty much from the ground up to overcome these limitations.

Submitted by Prof. Steven Sloman (Dept. of Cognitive and Linguistic Sciences)



Go Fish

We are currently setting up a behavioral test to test the visual acuity of Xenopus tadpoles and perhaps Zebrafish embryos as part of our lab's goal to understand early development of the visual system. We have a setup where we have a tank with a clear bottom that sits on a computer monitor. The tadpoles swim freely in this tank. We then play a moving stripe pattern in the bottom of the tank. Both tadpoles and fish exhibit an optomotor response, in which they swim in the direction of the moving stripes. By changing the size of the stripes we hope to test the detection limits of their visual system. While we are already working on developing the visual stimuli, we would like to be able to track an individual tadpole's movements throughout the duration of the test. To this end we plan to set up a computer camera to film them in the tank while the test is going on. We would like to have software that could then take these digitized images, identify the start location of the tadpoles and track individual tadpoles from frame to frame. The program would then output the x,y coordinates of each tadpole and compute various aspects of their swimming behavior, for example how much time they spent swimming, how much of their swimming was in the overall direction of the stripe motion, etc. Thus the project would have both an image analysis component and a data output and analysis component. It can be as simple or as fancy as the students wish to make it.

Submitted by Asst. Prof. Carlos Aizenman (Dept. of Neuroscience)



Patent Prober

One of the challenges in technology transfer is gathering information on a wide variety of subjects. When we receive an invention disclosure, we like to check issued patents and patent filings in the U.S. and abroad, publications by the inventors (inlcuding what they've posted on the internet), other publications in the field to get a sense of the state of the art, the names of companies (and prefererably individuals within the company) who might be interested in the technology and market information. We gather this by going to a number of different sources including some data base services to which we describe. It would be very helpful to have software that automatically pulls together such information.

An example in a completely different field is the Micro*scope project at MBL. Here they collect images and information on microorganisms from various sources and bring it together in a user friendly format.

I know of nothing that is available that is like this for invention assessment. Parts of it are available (eg. patent mapping tools) but nothing that is as complete as what I describe. Such a program would be of interest to academic technology transfer offices, venture capitalists, investors, law firms and high tech companies.

Submitted by Dr. David Kiszkiss (Brown Technology Partnerships)



Integration X

I want is a wiki-oriented intranet that provides calendaring and scheduling, a CMS for the dept web (to replacing SWP) and a secure and easily adaptable interface to a database of dept info (both system info and administrative info).

I planned to find the best oss projects out there and integrate them into a single website designed to be one-stop shopping for everything we do (facsearch, grad tracking, system configuration, office allocation, purchasing, grant administration, web-based small group interactions (e.g. courses, research groups) via wiki, website maintenance, etc, etc).

Submitted by John Bazik (Dept. of Computer Science)



Geoevents

I've often thought it would be great to be able to spatially visualize the upcoming events around Rhode Island. I'd love to see a system that is capable of parsing the Providence Journal's event page, make use of the Google Maps API, and represent the future events on a map. Some features that would be nice for the system to have include:

a) the ability to filter on certain types of events

b) the ability to step forward/backwards in time

c) the ability to input my home address, intially be presented with the distance to each event, and choosing a particular event should allow me to get driving directions

d) clicking on an event would provide all the details of that event

e) learning from my past events and providing a "suggestion" list would be a bonus

Submitted by Mark Dieterich (Dept. of Computer Science)



The Spatializer

This project would be a program that could more easily deal with the vast amounts of spatial data that is streaming from various continuous (remote sensing systems, environmental monitoring, economic indicators, traffic monitoring, etc.) and discrete (census, health care, surveys, etc.) sources. There are already tools for such applications. Perhaps a novel hook might be focusing on the visualization of such data for either sifting through the various types of data (metadata level), or for dealing with a single source of data (probably something publicly available, like weather or census).

Submitted by Asst. Prof. Scott Bell (Associate Director of Spatial Structures in the Social Sciences)



Facility Finder

The Facilities Management (FM) Department is interested in providing a web enabled solution to its constituency that will offer a user friendly interface into Brown's Facility information store. The project can be broken into different phases allowing the ability to scale the project up or down. The initial phase shall provide access to facility and project photographs. The FM Department currently houses thousands of facility photos, both interior and exterior, grounds, and project specific.

The solution shall provide the following features:

Web Interface:

- Photo display window

- Search feature using one or multiple meta data fields to select from.

- Search shall provide results as a page of thumbnails that the user can either continue to refine the results or scroll to and click to enlarge a thumbnail.

- Security permissions that allow certain end users to upload photos and enter associated meta data. Some fields will be mandatory fields. The ability to batch upload photos and have the photos automatically numbered. (i.e. Enter meta data on the first screen and upload a series of photos associated with one project.)

- Photo print and save capabilities, original size or scaled down thumbnail size.

Database:

Meta data for each photograph shall be held in tables within Oracle. Some lists, already exist in Oracle tables, such as Building List, other data tables are to be created. Photos will be stored separately from the meta data and fields will be required to point to the photo store file and location.

Security:

User access for the Brown Community will be determined by the users netAuth ID. (i.e. Student, Faculty, or Staff). Non-Brown access will need to be determined by Facilities staff and require an Administrative interface to create and manage login accounts. Members of certain departments will need access to certain project or building related photos as determined by FM staff. (i.e. A department coordinator will have access to the detailed progress photos of a renovation project being done for that department.)

The general public will have access to standard generic photos as determined by FM staff. These photos can be tracked with meta data as well. (i.e. General users will have access to Exterior Building photos.)

Submitted by Monty Combs (Facilities Management Department)



SQUAT (Sequence Quality Assessment Tool)

Background: Highly active anti-retroviral therapy (HAART) for HIV/AIDS has been available in western countries in the past 10 years and has recently become available in resource-limited setting where the majority of the HIV epidemic exists. Drug resistance is the main cause of HAART failure, and its monitoring is crucial for the success of initial and subsequent therapy. This monitoring includes genetic sequencing of HIV drug targets - protease and reverse transcriptase (RT) - and determining presence/absence of mutations that confer resistance. Global access to antiretroviral therapy is increasing and as a result, genetic sequencing is expected to escalate as well, and quality-assessment tools are essential.

Methods: We have developed a sequence quality assessment tool (SQUAT) that can be used for quality control prior to sequence data analyses. The tool combines Perl and R scripts and is run in the R environment. Cut-offs for sequence analyses are calculated from a large HIV sequence dataset (http://hivdb.stanford.edu). Genetic distances are calculated using SynScan (http://hivdb.stanford.edu/pages/synscan.html).

Results: Nucleic acid protease and/or RT sequences are read into SQUAT either individually or in batches from user-submitted files, are aligned and translated, and are screened at the nucleic and amino acid levels for frame shifts, insertions, deletions, stop codons, ambiguous characters, and atypical mutations. Lastly, pairwise genetic distances are measured to identify outliers. The SQUAT output includes a summary report containing approved and non-approved sequences with detailed information based on the screening categories, a histogram of pairwise genetic distances with phylogenetic tree, and amino acid sequence files for subsequent analyses.

Conclusion: SQUAT is a stand-alone tool designed to ensure appropriate sequence quality and to facilitate troubleshooting HIV protease and RT sequences subsequently used for drug resistance identification.

The purpose of this project would be to turn this tool into a piece of software that is easily installed and has a nice front end. An additional possibility is creating a website for future updates and potentially automatic updates from various sites that 'feed' SQUAT.

Submitted by Allison DeLong, Joe Hogan, and Rami Kantor (Center for Statistical Sciences)

We are currently looking for new project ideas for CS 190. If you have any ideas for potential projects, please e-mail theCS 190 TAs to have a project added to the list.

For examples of previous projects, you can look at Project Ideas From 2005

Last modified: Fri Jan 20 16:46:01 EST 2006