Daniel Querales
Data Scientist
github: github/dquerales linkedin: in/danielquerales
Electrical Engineer with Data Science expertise. Proficient in R, Python, and SQL. Experienced in all stages of the machine learning lifecycle, including data cleaning and processing to model deployment and monitoring. Knowledge of continuous integration and continuous delivery (CI/CD), version control with Git, and cloud computing platforms like Azure, GCP, and AWS.
Projects
Two year of COVID-19 - View on GitHub
  • Interpreted successfully the data from the COVID-19 dataset to explore the development of the pandemic.
  • Utilized MySQL to identify the values of total cases, total deaths & people fully vaccinated.
  • Built a dashboard in Tableau Public to report findings, including key metrics.
 
Predicting Tesla Stock Prices - View on GitHub
  • Predicted Tesla (TSLA) stock historic prices using a linear regression model to identify trends.
  • Explored prices (adjusted close, volume, market cap, volatility) with Python tools like Pandas, Numpy and Plotly.
  • Achieved 98% in R2-Score using the model with test values.
 
Sales Analysis: Electronics Store - View on GitHub
  • Performed an analysis of the sales records from a retail to discover opportunities to increase revenues.
  • Merged multiple files into a single CSV, transformation from multiples dates formats and data cleaning operations.
  • Extracted insights about the best products sales, the best months for sales, products often sell together and most frequent hour for sales.
 
Customer Segmentation: Clustering - View on GitHub
  • A unsupervised machine learning project for clustering customers for a marketing campaign.
  • Data cleaning (like outliers, nulls), preparation and transformation of a customer dataset. Also exploration and visualization of features.
  • Used Kmeans clustering with WCSS and Silhouette averages for scoring.
 
Fraud Detection for Bank Transactions - View on GitHub
  • A logistic regression project for fraud detection in finance environments.
  • Used classic, gaussianNB, svc and random forest models for detection.
  • Assessed performance with accuracy, precision and recall scores. Plot ROC curves to visualize AUC (area under curve).
 
Product Recommendations for Retails - View on GitHub
  • An apriori algorithm project for product recommendations for a retail store.
  • Queries for most popular items, best month for sales and products often sold together.
  • Used Apriori for mining frequent products sets and relevant association rules to make products recommendations.
 
AutoML: Automated Machine Learning Workflow - View on Streamlit
  • An automated machine learning workflow for classifiers models.
  • Created a preview from the uploaded data using pandas profiling.
  • Prepared, cleaned and trained the model using AutoML Pycaret
 
Web APP deployment: Sentiment Analysis - View on Streamlit
  • Developed a text classifier model for sentiment analysis of movies reviews to reduce human intervention.
  • Vectorized the text reviews for their use by the SVC classifier model, then created a pipeline to automate the input prepossessing to increase accuracy performance.
  • Deployed the model as an Streamlit app for their use.
 
LLM Agent: Using CHAT-GPT for data exploration - View on Streamlit
  • An large language model agent for data exploration.
  • Used chat-gpt as base llm model.
  • Set the agent using langchain libraries
Work Experience
Data Scientist | 2022-Present
QUASH.ai
Entry-level Data Scientist
 
Freelancer | 2019-2022
Freelance
Data Science, Hardware Description, Numerical Methods and Control Systems
Education
Engineer’s Degree in Electrical Engineering | 2008-2019
Universidad Central de Venezuela
Electronics, Computer Science and Control Systems