Stock and Crypto Forecast with ML and DL
Deployed best models with Streamlit API.
The objective of this platform that I created with Sebastian Esponda, and deployed with the API through Streamlit, is to help investors and users to have a competitive digital advantage, reduce losses, complement the analysis with valuable information on predictions and data visualizations, and improve their investment portfolios.
Results
+20
ML and DL models trained and tested
1
Interactive Model Deployed
90%
Grade on the Final Masters Project
Goals
We have structured this data science project using the CRISP-DM methodology (Cross Industry Standard Process for Data Mining) which consists of 6 steps. (Annex 1: CRISP-DM Methodology) This process begins with the Business Understanding, understanding the environment, business needs and problems, and ends with the deployment or "Deployment" of the model developed for its final use by the user.
The objective of the work, explained in a more technical way, is to make a series of predictive models on a set of stock and cryptocurrency data, in order to predict the prices of these financial assets through an application deployed on the Internet. We have developed these models and applications with a series of libraries, APIs and codes, such as NeuralProphet from Facebook, Pycaret and Keras from Google for the development of predictive models with neural networks, Yahoo Finance to import financial data from stocks and cryptocurrencies (a through scrapping), Plotly to visualize results, and others that we have attached in the Annex (Annex 2: Libraries, Code and Resources) and in the Notebook with the code of all the work.
To carry out the exploratory analysis and models, Google has been chosen as the reference stock. On the other hand, in the API you can choose more financial assets such as Bitcoin, Apple, Amazon, etc.
Deployment for Stock and Crypto Forecast Models: Play with it!
Deliverables
Colab and HTML code files. Composed by:
- Machine Learning Models
- Deep Learning Models
- EDA and Feature Engineering
8
Website for Deployment
1