CSI RESUME ANALYZER

Yalvac Top

Ali Kemal Tanriverdi

Elcin Can Cavusoglu

Content

• CSI Resume Analyzer Assumptions
• CSI Resume Application Flow Chart
• How do we calculate grading?
• Resume acceptance criterias
• Current state of the project
• Next iterations

CSI Resume Analyzer Assumptions

• Need an active internet connection on the operating machine and the device which takes pictures of resumes
• Need to have a dropbox or google drive account which synchronizes the phone and the machine where the code will execute on
• The input should be either pdf or any kind of image type (.png, .jpeg, .jpg)
• The program needs to be active when the user takes pictures for prioritizing resumes

CSI Resume Application Flow Chart

Name Recognition Algorithm & Association with the candidate

Resume Analyzer Algorithm

Key Words

Result Prioritized Resumes

Store Image & Text files

Image to Text Converter

How do we calculate grading?

• Analysis modules search through the text file for negative keywords. If any match is encountered, then the grade of the resume is decreased by 1 point, and the reason for point reduction is added back to the database to the reason column
• Analysis modules search through the text file for positive keywords. If any match is encountered, then the grade of the resume is increased by 1 point, and the reason for the last state of the points is added back to the database to the reason column

Resume acceptance criterias

• If the grade of a resume is above 0(zero), then resume is added to the accepted resumes table in the database with its reasons. It is placed at the top section of the application window
• If the grade of a resume is equal to 0(zero), then resume is added to the waiting resumes table with its reasons. It is placed at the middle section of the application window
• If the grade of a resume is less than 0(zero), then resume is added to the declined resumes table with its reasons. It is placed at the bottom section of the application window

Current state of the project

• Able to successfully transform any picture or .pdf to any desired format by using google vision API to text format
• Stores, updates and retrieves them at the database when needed
• Able to do a primitive analysis on the uploaded resume to the database according to the given user preferences
• Demo

​

​

​

Demo

All technologies we are using

• DropBox or Google Drive for the synchronization between the smartphone and the system
• OCR which is provided by Google’s Vision API, detects the letters and sentences in an image and transforms them into text
• Program is implemented using Python because we found more resources and documentations
• We are using MySQL as database and Valentina Studio database management application tool

Biggest Challenges

• Prioritizing resumes according to the analysis by using good and bad words
• Increasing the name extraction rate of name recognition algorithm: Currently, we can detect 80% of the names correct from 1 page resumes
• Finding gender and race of the candidate
• Determining age, work ethics, characteristics of the candidate by investigating social media accounts
• UX of the user interface

Next iterations

• Increase percentage of successful name extraction of the candidate from the resume by using name entity recognition algorithms
• Develop analyzing algorithm to meet the needs of the user
• Create a develop analysis of the candidate not just by using the information given but by analyzing data provided on social media or other platforms
• Develop database design to meet the needs of the user
• Get the input and return the results through an app developed for the user mobile device to avoid using third party software like dropbox or google drive

​

​