I am a Mechanical and Electrical Engineer ⚙ from Tec de Monterrey
and currently pursuing a Master's degree in Data Science 🧑🏽💻 at ITAM.
Before I started my path as a Data Scientist, I worked in the USA as a Field Service Engineer at SEAM GROUP, performing infrared thermal inspections on live electrical connections,
I learned to focus on the tiny details to prevent failures. I also worked for two years as a Service Sales Engineer at Grundfos,
selling predictive maintenance contracts and other service solutions in the Central and SouthEast area of Mexico. During this experience,
I learned to identify and solve customers' needs by having effective communication, increasing 30% of active contracts in 2020.
- 🌱 I am currently learning about Deep Learning and Bayesian Statistics.
- 📫 How to reach me: jh.escalona.s@gmail.com
Recent Work
Nayarit Shelter Locator
In this project, I created a simple Shiny dashboard to locate the nearest shelter in case of an emergency in the state of Nayarit, Mexico.
First, I developed an ETL to clean the raw data from a Google Sheet to use it. The dashboard has three main windows;
the first one is a table that shows the shelter's information so that it can be filtered by the municipality,
address, and type of building. In the second one, there is a map that shows the location of each shelter, by clicking
on the pins info is displayed, it can also be filtered by the municipality. The last window helps you find the nearest
shelter given a location.
Creating a Multidimensional Financial Index using PCA
In this project, I created a Multidimensional Financial Index using PCA of the most important financial market in the world, the North American one,
in a AWS EC2 instance. In particular, the historical data of the last four years was obtained from the stock price of Dow Jones Utility, DJ Composite,
CBOE Volatility among others. Using the power method to obtain the maximum modulus eigenvalue with its respective eigenvector, and the deflation
method to obtain the rest of eigenvalues and eigenvectors.
Art Classification
In this project, I trained a CNN (Convolutional Neural Network) to classify different types of Art from WikiArt,
Visual Art Encyclopedia. The data was obtained by scraping the WikiArt page with Python. Four styles of Art were selected to classify, the first approach
was to create a CNN from scratch, but the results were not as expected. Therefore it was use transfer learning, using as base a pre-trained model,
in particular DenseNet121, and adding other layers to improve performance.
Get In Touch
If you want to know more about me, feel free to contact me.
