unCommon CV

Inspiration

One of our founding members in her research at the University of Waterloo had to brave the Common CV--a monumental form required to be completed in the research grant application process. Our other members have also heard of this infamous form and how tedious and time consuming it is, and decided to create a solution to automate, and thus expedite, the process.

What it does

The user uploads their Google Scholar profile to our website, unCommon CV, and then the script we wrote recovers a table containing all important information of the papers the user authored. This includes the title, DOI, authors, date, summary, and many more important features. No extraneous work is required to search for the citations belonging to all the papers an individual has wrote--with a simple click, process is several orders of magnitude faster.

How we built it

Initially we identified a solution that included Machine Learning algorithms and complete automation. However, once we understood our constraints and requirements better, the team shifted away from design complexity, removing time-consuming features that were not necessary for the MVP.

The three main application modules are 1) Data mining and processing module 2) Back-end layer, and 3) Interfacing module. We kept the back-end layer as thin as possible in order to save cost and development time. The work was equally split between all three team members working on different aspects of the system.

Challenges we ran into

As we had expected, interfacing our application with the Common CV's was the most challenging part. Common CV lacks API or any other programmatic interfaces. Despite automation being the team's top priority, we decided to sacrifice some of the automation in order to complete the project on time.

Accomplishments that we're proud of

From what we had learned over the course of the past 36 hours, the most extraordinary feat for us was being able to make the front-end compliment the back-end work. Individually, we had all grown massively in our own computer science comprehension, and for a first Hack The North experience, it without a doubt served as a memorable.

What we learned

For Kelly, she had learned how to further her own learnings in python by using the Web.py and Beautiful Soup libraries, not to mention picking up PHP when working on the development of the back-end script. For Tim, he had a masterful comprehension of Python, and discovered new errors to troubleshoot when working with establishing the server for our script, and worked to interface the code with the front-end website. Sahad had picked up HTMK and CSS to lay the framework for the website, and then took the next step to increase the site's aestheticism with Bootstrap. Him and Tim were able to create a beautiful website. Tim also took a leadership role in the code development, providing guiding questions and tips to both Kelly and Sahad in finding a long-overdue solution to this problem.

What's next for unCommon CV

The next step would be to allow an option to upload a resume to unCommon, in which the website would fill CCV areas associated with information on the resume. We had tried to get a script going to parse over an uploaded resume and recognize/pull out data such as names, addresses, education, and experience. We had initially used the docx library from Python, but this proved to be a fairly large challenge in terms of examining all the possibilities of what could constitute important information. This would be the next step in terms of improving our product. Additionally, machine learning could possibly be run to improve the new script's ability to discern what is important data from the resume.