Isabela Ruiz Roque da Silva

1 ISABELA RUIZ ROQUE DA SILVA 26 years – Brazilian – Single Sao Paulo – SP + 55 -- ACADEMIC EDUCATION § Motion Graphics Designer – After Effects AXIS (School of Visual Effects) – Sao Paulo, SP Conclusion: JUN/2020 § Academic Program for Special Effects (VFX) – Maya, Nuke, Houdini AXIS (School of Visual Effects) – Sao Paulo, SP Conclusion: FEB/2020 § PhD in Electrical and Computer Science – Blockchain, IOTA and cryptocurrencies Mackenzie University – Sao Paulo, SP Conclusion: DEC/2022 § Master in Electrical and Computer Science – Data Science, NLP, Deep Learning, Classification Algorithms Mackenzie University – Sao Paulo, SP Conclusion: JUN/2018 § Computer Science College of Computer Science and Business Administration (FIAP) – Sao Paulo, SP Conclusion: DEC/2015 § Computer Science – High School Madre Mazzarello – Sao Paulo, SP Conclusion: DEC/2010 LANGUAGES § § § § English: Fluent Spanish: Advanced Mandarin: Basic Japanese: Basic PROFESSIONAL EXPERIENCE § Mackenzie University Laboratory Worker (SEPT/2019 until now) Main Activities: Coordinate interns inside the software factory and development of PhD research with Blockchain, Tangle and Cryptocurrencies. § GPA Data Scientist (MAY/2019 until AUGUST/2019) Main Activities: Price optimization models to find the best price for all products with R. 2 § Duratex Data Scientist (NOVEMBER/2018 until MAY/2019) Main Activities: Sales Forecasts for the SOP Area, Customer Segmentation for analyzing clients and potential clients and System Recommendations with R and Python. Some libraries used: - TBATS; - Holt Winters; - Neural Networks; - VAR/VEC; - Pandas § FCamara Consulting & Training (Duratex) Junior Data Scientist Consultant at Duratex (JULY/2018 until NOVEMBER/2018) Main Activities: Creating Sales Forecast using statistical models validated by the Academy. § J.P. Morgan Chase Internship in Corporate Investment Bank Technology - Systems Architecture (APRIL/2014 until DECEMBER/2015) Participation in regional projects to migrate legacy systems to more modern and secure applications, using the Pentaho Data Integration tool. Main Activities: Automation of processes based on user specifications, direct contact with other areas of Technology and Business, conducting project tests together with users. CERTIFICATES AND AWARDS • • • • FCE – Cambridge (English) HSK Level 1 (Mandarin) 1st place on a workshop competition with astronaut Dan Barry in programming small helicopters using Arduino. 2nd place in the Spacecup competition: modeling, 3D printing and launching of a rocket to reach a specific target. PROJECTS • Deep Learning applied to Classification of Emotions in Social Media Texts Emotions are important in interpersonal relationships and are part of the daily life. Many people today express themselves through texts and posts on social media, such as Twitter and Facebook. This research proposes the use of Deep Learning techniques to perform the classification of emotions in these texts, using the Ekman’s basic emotions model: anger; sadness; fear; disgusted; happiness and surprise. To create the input matrix of the deep neural network, we applied the Word2Vec model, which has two learning algorithms: CBOW and Skip-gram. • Analysis of Dengue fever spread in Brazil using Twitter and Big Data This project presents a detailed study of Dengue fever spread in Brazil using Twitter, Big Data, R programming language, with tweets on a NoSQL database called MongoDB. As a conclusion, the information that the users share on Twitter claiming to be with Dengue fever were compared with official sources of the Ministry of Health and it was possible to verify the same trend. SKILLS • EDITING TOOLS 3 PhotoShop, Dreamweaver, Flash, Corel Paint Shop, Sony Vegas, Ulead Video Studio, Adobe Premiere. • MOTION GRAPHICS/VFX Adobe After Effects, Autodesk Maya, Nuke. • MACHINE LEARNING Neural Networks, Deep Learning, Supervised Learning, Natural Computing. • PROGRAMMING LANGUAGES Python, C, PHP, R. • WEB DEVELOPMENT HTML, CSS, Javascript. PAPERS • Classifying Emotions on Twitter Messages using a Deep Neural Network Paper presented at DCAI 2018 Conference, Special Session on Web and Social Media Mining (WASMM). Many people use social media nowadays to express their emotions or opinions about something. This article proposes the use of a deep learning network architecture for the classification of emotions in Twitter messages, using Ekman's six emotions model: happiness, sadness, anger, fear, repulsion and surprise. We collected tweets from a labeled data set that contains about 2.5 million tweets and use the predictive model Word2Vec to learn the relationships of each word and turn them into numbers that the deep network receives as input. Our approach achieved accuracy of 63% with all classes and 77% accuracy in a binary classification scheme.