CE-19.09

Title: 

Web Application for Woman Owned Business

Author(s): 

Mrinalini Garre

Oral Defence Date: 

Monday, August 12, 2019 - 13:00

Location: 

TH 434

Committee: 

Prof. Ilmi Yoon and Assist. Prof. Abeer AlJarrah

Abstract: 

Today, Women entrepreneurs around the world are making a difference, they are growing in numbers and creating increased socio‐economic potential, in terms of greater social inclusion, revenue and technological innovation while also reducing inequality in society. However, women’s entrepreneurship and business ownership are often ignored in comparison to that of men. A recent study conducted by the PitchBook data shows that in 2017, the woman entrepreneurs received only 2.2% of all venture capital in the US. As a solution to the problem, companies or government units that want to provide special funding or training to women owned companies manually search and validate Woman-owned companies. In this project, we develop an Automated Gender Predictor web application which identifies woman owned businesses. The web application is developed using Python and Flask framework, where the user uploads a secretary of state approved PDF document from which we extract information such as name and profile picture of the business owner by scraping. We used FreeOCR API to perform PDF scraping to extract name of the owner and Selenium Webdriver tool for scraping the Facebook to extract the profile picture of the owner. The obtained information which is first name and facial image is used for predicting the gender of the owner by implementing Machine Learning model. To achieve this, we built an LSTM model using Recurrent Neural Network to predict gender from the first name and VGG neural network model using the Convolutional Neural Network to predict the gender of the owner from the facial image. The results obtained from the machine learning models predicts the gender of the person which helps in identifying the woman owned business.

Keywords: 

Woman-Owned Business, Gender Detection, Machine Learning, Scraping, Keras, LSTM, VGG Neural Network.

Copyright: 

Mrinalini Garre