Make a website for this club! We're looking for a student leader to help develop and maintain the web applications that drive Inve[n|s]tUB.
There are a plethora of social media sites that make up our online networks. While each may have a personal niche in the way that we use it, there are times when it would be great if we could post to all of our networks simultaneously. The idea would be to develop a service that lets you configure multiple social media accounts (twitter, facebook, linkedin, google+, etc) and have a single post go out to all of them via API calls.
Create an extraction package and data visualization framework to load daily data updates from the Wharton financial data set.
Develop an asset management system for hardware in the undergraduate CSE labs. This will let students know what is available when, if they need specialized tools like echo, hololens, etc.
On Board Diagnostics Reader Photo Detection Goal: Develop an automated way to determine which photos from a set of vehicle photos contain an OBDII reader and its on-screen output Description: Unfortunately, rugged OBDII readers that integrate well with mobile devices or cloud APIs are lacking, and OBDII information is currently included on the ACV Auctions platforms via photos of OBDII reader screens. We would like to be able to identify which photos in a set of vehicle photos are photos of the on-screen output of an OBDII reader. This may require some dataset development depending on problem solution developed. Skills: machine learning, computer vision, general coding skills
On Board Diagnostics Text Extraction Goal: Develop text extraction from the photo of the screen of an OBDII reader and organize the extracted text information in a usable format Description: Unfortunately, rugged OBDII readers that integrate with mobile devices or cloud APIs are lacking, and OBDII information is currently included on the ACV Auctions platforms via photos of OBDII reader screens. We would like to be able to read the text displayed on-screen by OBDII readers, and aggregate this in a more useful format. This information could be included in our Condition Report provided to users on our platform, and codes extracted could be matched with related issues. Skills: machine learning, computer vision, general coding skills
Tire information extraction from tire sidewall Goal: Extract tire information from photos of sidewall Description: Tire information such as make, brand, and condition are factors that can contribute to the determination of a vehicle’s value. By means of computer vision and an iPhone XS factors will be extracted and translated from the tire sidewall to a mobile app. Bonus points will be awarded for determining tread depth and dry rot.
Website/ Mobile Application: Willo Willo's vision is to simplify the will-making process and provide a simple platform that allows anyone to comfortably and conveniently create and maintain their will. 78% of millennials (ages 18-36) do not currently have a will, meaning there is no plan in place for where their assets will go if a tragedy were to occur. These are also the same people who are making big life decisions, such as career choices, starting families, buying homes, making investments, traveling and more; making a will more important than ever. The current method of creating a will is a tiresome and lethargic process, but it doesn't have to be that way. New generations are already equipped with the skills necessary to operate computers and mobile devices, while older generations are continually adapting to this ever-increasing trend. With powerful technology in the palm of our hands, Willo delivers a service that is universally user-friendly and allows anyone to create and update their will anywhere, anytime, right from a mobile device or computer. With Willo we can reshape the way the world views will-creation and maintenance.
Create a prototype system for automating transcription of audio recording with subsequent natural language processing. It would be a pilot exploration…two examples of use would include analysis of job postings (obviously this is only NLP focused) and transcribing a CAD-based design exercise.
Develop an app for UB’s award-winning student newspaper, The Spectrum. The app should link to our content and allow students to: 1) Receive warnings of emergencies, breaking news, closures etc. 2) Send news tips or info about a developing story to a “hotline” 3) Submit an image and tagline from around campus (with photo credit to the student sender) 4) See events happening around campus/Buffalo in the upcoming week (possibly month)
ACV Auctions, the fastest growing online wholesale automotive marketplace, sells over 20,000 cars per month all across the continental United States. The property title for a vehicle must be verified and processed for each car sold. This is a labor intensive, tedious, and error prone process. In addition to the title document, various other documents may be attached such as notary, lien release, marriage/birth/death certificate, etc. which may affect its processing, and odometer readings and other items may be handwritten. The document itself varies greatly in structure from state to state, i.e., a New York title may be vastly different from a Texas title (and we sell a lot of cars in both!). We would like industrious students to assist us in developing an impactful solution to the challenge of title processing. The project will involve various computer vision problems such as document scanning, classification, text recognition, and optical character recognition. Excitingly, most of these components are relatively well-solved problems, e.g. OCR, where there are many out-of-the-box and even open source solutions. We even have a base proof of concept app for you to start with! We anticipate a potential backend (Python) and expanded frontend (iOS) in the solution and look forward to helping students develop skills in computer vision, deep learning, Python, Android and iOS development!
Create an ap to help bring financial awareness to the younger generation