All too often have heard from people close to me that they have submitted so many job applications but have not heard from any or just a couple of out of say 20 applications got responded to. And even after that, getting an interview is less likely. Earlier in my career I too had some similar experiences. This is what triggered this project. Although academic it gives an individual some sense of his or her chances of getting an interview when applying for a job.
How it works
The candidate needs to answer several questions using application. Once these are provided and the candidate hits the submit button, the engine weighs each response and then using a basic algorithm produces a probability value. This is displayed to the candidate.
Challenges I ran into
The data to support the algorithm is not available at all or is only available with organizations that collect such recruiting data.
Accomplishments that I'm proud of
This was my first attempt at producing a user-facing application using R and Shiny. It was also a fairly new concept which I did not see on the web so far. I am also happy with the fact that despite the of lack of real-world data, I was able to model the problem using personal experiences, some feedback from colleagues and then translate these into a based set of generated data.
What I learned
Data is key to solving such predictive problems. Always start with a problem to solve and the solution, technologies and building blocks will fall in place.
What's next for predinter
Enhance this with live data. Tweak the questionnaire to domain-specific use case and also to general ones. Associate a feedback loop into this. The algorithm is simple enough and does not need much tweaking, but the user interface can be enhanced further.