This project is focused on practicing and integrating machine learning techniques into a Django application. It involves exploratory data analysis, model selection, and hyperparameter tuning, as well as creating a predictive system using train test split. Additionally, the project covers web scraping, Django forms and models, file input, CRUD operations and the use of charts in a Django application.
The application aims to predict whether a bank client is likely to take a term deposit when contacted.
The jupyter notebook file with the machine learning model generation is included in the GitHub repository.

CapitalCall

This project focuses on the practice and integration of Deep Learning techniques to classify emails as spam or ham. It includes exploratory data analysis, model selection and tuning, and the creation of a predictive system using training and validation sets. Emphasis is placed on text preprocessing and feature extraction using pre-trained embeddings. The implemented neural network architecture combines an embedding layer using GloVe vector representations, Conv1D convolutional layers, and bidirectional LSTM layers, along with regularization techniques such as Dropout and L2 to improve generalization.
Subsequently, an analysis of the incorrectly classified emails is performed.

Spam Email Detection

This project focuses on practicing the use of Machine Learning to perform customer segmentation using clustering. Some of the skills and techniques used in this project include: exploratory data analysis, working with dates, using the K-Means model for clustering, using One Hot Coding for categorical columns, the "Elbow" method for choosing the number of clusters, dimensionality reduction using T-SNE, and analyzing each cluster using mean, mode, and other indicators.

Customer Segmentation

This is the code of this website. Created in NextJS, it uses a MongoDB data base to storage the projects information.

Portfolio - NextJS

Django application that uses a Logistic Regression model (Machine Learning) to predict whether a job offer is fake. The repository includes the Jupyter Notebook file where the model was fit.

Fake Job Detector

Django app that allows saving passwords in a MySQL database. There is an authentication system with a custom user (the default Django user was modified). CRUD operations can be performed on the passwords and there are search bars for "name" or "type" which are defined by the user.
The app also has Docker files that allow running it on containers using Docker Compose.

Password Vault

This was the final proyect at 4Geeks Academy. It's an app that allows users to put their pets for adoption, receiving applications from other users and choose based on their responses to an adoption form. It also allows to send adoption forms to be a candidate to adopt another user's pet. 
The front-end was made with React, while the back-end is based on Flask and SQLAlchemy.

Appdopta

Second project done in a self-taught manner. Consists in a simple HTML interface which allows entering input data for a dehydration process. With those, the mass and energy balance are calculated and some relevant process parameters are shown, like drying time, gas consumption, etc.
This project is very OOP oriented; each equipment and mass flow is represented by an object with their respective attributes and methods.

Dehydration Plant

Discord bot for the game South Park Phone Destroyer, it has commands for the game and other commands to interact with other users from the discord server.
This bot is written in Node, uses DiscordJS to register slash commands and Express to turn it into a web application which allows keeping the bot active using the website uptimerobot.com.

LuckySloth - Discord Bot

Initial version of the discord bot for the game South Park Phone Destroyer, it has commands for the game and other commands to interact with other users from the discord server.
This bot is written in Python, uses Pandas to calculate things related to the game and Flask to turn it into a web application which allows keeping the bot active using the website uptimerobot.com.

Blanket - Discord Bot

Proyect for 4Geeks Academy. It's an app that consumes an external API and shows relevant content using React. The Rick and Morty API shows the information paginated, so it was added an option to choose the page. For each character you can click Learn More and shows detailed information (also obtained from the API).
The front-end is based on React, HTML and CSS. There is no back-end.

Rick And Morty Blog

Project from 4Geeks Academy. It's a music player built in React. The audio files are obtained from a list of links and they are played using the Audio() constructor.

Slothify

Project from 4Geeks Academy. Game Tic Tac Toe created using React. Both players write their names and the player 1 picks a symbol at the start. The players take turns for their plays.

Tic Tac Toe

Project from 4Geeks Academy. Game Battle Ship created using NextJS. It's one player against the computer. Ship positions and computer shots are random.

Proyectos

Filtros

CapitalCall

Spam Email Detection

Segmentación Clientes

Portafolio - NextJS