Serene-1.0

Fiels: Data scraping, data analysis, NLP machine learning

Technologies: Pushshift.io, Python, Keras with TensorFlow, Scikit-learn, Google's Colab

1. Description of the project
2. Initial project ideas
3. Accomplishments

Description of the project

Serene is a web based application that implements a machine learning model meant to detect depressive or non-depressive tendencies in the user's texts. This is a project that I intend to further develop and transform it into a reliable tool for people.

Initial project ideas

• Scrap data from Reddit.
• Preprocess the data.
• Build NLP models for text classification.

Accomplishments

• Scrap data from Reddit's subreddits with Pushshift.io API.
• Preprocess the textual data.
• Build the machine learning models, using Scikit-learn and Keras with Tensorflow.
• Develop a deeper knowledge of Logistic Regression, SGDClassifiers, Word-level CNNs, Character-level CNNs and the optimization of deep neural networks.
• Train the models on Google's Colab Jupyter Notebooks, using their backend computations.