Sentimeter

An automated sentiment measurement tool and summarizer

Comment

Inspiration

Currently evaluating candidates is a time consuming process. While evaluations are still something that has to be done by people, we want to streamline and automate the process of consolidating multiple evaluations and generating an analysis of the evaluations.

What it does

To solve the problem of summarizing and analyzing the evaluations of candidates, we will first use the Smmry.com API to summarize multiple reviews, and then we will use the natural language understanding APIs of IBM Watson to perform sentiment analysis on the summarized evaluation. The API provides analysis of five base emotional responses in the evaluations: Joy, Fear, Anger, Disgust, Sadness. This analysis will be provided on the dashboard to provide a quick summary of the candidate.

How we built it

We build a front end web site with HTML5, CSS3, Bootstrap and Javascript to handle the input of candidate evaluation and the output of sentiment analysis and display into charts. We used PHP code to call the Smmry and Watson APIs

Challenges we ran into

We started off with the main idea of helping out the HR department in their struggle of sorting and selecting the best applicants along with the possibility of providing some automated feedback to the applicants that are concerned. The major challenge was to discover the problems that managers are facing on a day-to-day basis. Some of the problems that we were able to identify: Requirement of a summary or a brief write-up pertaining to the interview of a candidate and some sentiment analysis of how the entire interview went.

Accomplishments that we're proud of

We were able to successfully deploy a working model of the sentiment analysis and summarization functionality. We have also made efforts to further our progress in the field of resume parsing.

What we learned

We have utilized a bunch of really good API calls for this project to be a huge success. We started off with the Smmry API for the summary generation of feedbacks given from the HR dept, Technical Recruiter and the Technical Team. Later, we went for using IBM Watson's Natural Language Understanding API for the sentiment confidence score and the emotion sensing.

What's next for Sentimeter

We anticipate that our tool can be used in a number of scenarios. Employers may wish to use this tool to evaluate candidates and verify that the hiring manager’s decisions line up with the automated analysis in order to measure the effectiveness of the hiring process. They can also use this to train new employees in their HR departments by using past evaluations and analyses as a training tool. This tool can also be used as part of a mock interview to help recruiters train candidates to perform better during the interview process. Also for candidates to evaluate themselves or screen cover letters, emails, etc. Further we feel the application can also be utilized for Resume Parsing and extraction of data from Twitter, Facebook and LinkedIn to dig deeper into the candidate's social media aspects.

Share this project:

Updates