Laura Urdapilleta

About Laura Urdapilleta

As an AI specialist with a passion for driving business growth, I thrive on working with clients in diverse industries to identify and implement effective solutions using the power of Artificial Intelligence and Cloud Computing. I also value describing technical concepts to a non-technical audience. I bring a unique combination of academic background in evaluation, measurement, and research design, with experience in data science and data management. I am passionate about ensuring data accuracy and integrity, engaging stakeholders effectively, and assessing the impact of programs and I welcome opportunities to serve with my continuously developed core technical skills: research design, metrics development, artificial intelligence, machine learning, business intelligence, data management, data wrangling, and program evaluation. I cherish opportunities for public speaking and team collaboration. I especially enjoy incorporating Agile frameworks like Scrum into project management. Contact me if you are interested in building analytics solutions, evaluating a strategy\'s effectiveness, or assessing the validity of your metrics. Let\'s connect: l••••••••@gmail.com Who am I? I am a mentor. An optimist. An Arts lover. A listener. A communicator. SKILLS: • Metrics creation, variables operationalization • Regression analysis: linear, logistic, non-linear, multiple, ordinal, polynomial, lasso, ridge • Hypothesis testing, A/B testing • Analysis of Variance: ANOVA, MANOVA, ANCOVA, Chi-Square • Missing values analysis • Feature engineering • Multicollinearity, outlier, and residual analysis • Supervised and unsupervised classification • Natural Language Processing: Topic Modeling, Sentiment Analysis • Web Scraping • Sampling • Power Analysis • Data management and integration QUERYING AND SCRIPTING LANGUAGES: • Python • SAS • SQL DATA VISUALIZATION: • Tableau • Power BI • Seaborn • Matplotlib OTHER LIBRARIES: • Pandas • Numpy • Scikit-Learn • NLTK • Spacy • Beautiful Soup • Selenium • Tensorflow • Keras

Experience

1. Data Scientist

   Metis

   Aug 2022 — Present · Chicago, IL, US

   Completed a 16-week data science boot camp and designed five end-to-end projects covering data acquisition, mining, model fitting, evaluation, and communicating results using various visualizations. Projects included: 1): • Developed a PyTorch-based deep learning Convolutional Neural Network (CNN) using OpenCV to classify 4,105 paintings by emotional response for Art Therapy. • Used transfer learning (ResNet-101) and data augmentation to address class imbalance. • Achieved a 15% increase in user engagement through personalized recommendations. 2): • Developed a supervised learning model to classify seismic damage in buildings, assisting Nepal\'s National Reconstruction Authority in selecting low-damage proposals. • Trained and validated the model with post-disaster datasets of 260,601 buildings, using resampling methods to address class imbalance and evaluating multiple algorithms. • Model integration into bidding evaluation could impact 4,106 lives and prevent US$3.5 billion in losses during an earthquake in an urban area like Kathmandu, Nepal. 3): • Analyzed movie reviews using NLP techniques (Topic Modeling) to identify 7 critical topics related to audience preferences by genre. • Preprocessed data with Spacy and employed NMF supported by tf-idf to differentiate the main topics. • Visualized results using Tableau, helping a movie production company determine what movies to produce, market, or distribute based on audience preferences and expectations.

Education

• Western Michigan University

  Ph.D., Evaluation, Measurement and Research

• Western Michigan University

  Graduate Certification, Applied Statistics