Web App: https://share.streamlit.io/esdfirls/wine-prediction/main/streamtest.py
The objective of this project is to predict the quality of a wine from its physicochemical properties and deploy the solution in a API, Desktop App and Web App. Data were extracted from kaggle for study purposes. There are several possibilities of application in development and quality control in the wine making process from this solution.
Based on the properties of the wine, a prediction is made whether the wine is good or bad. The database has the properties and an evaluation score, which goes from 0 to 10. These groups were separated where 0 to 5 is bad and 6 to 10 is good.
Problem: Identify the quality of a wine from its properties
Gaps: A taster's analysis can be subjective and may not be sufficient to determine the quality of the wine.
Hypothesis: The analysis of physicochemical properties can be used to create a model with the objective of predicting the quality
Goals: Extract data from wine data, analyze and apply Machine Learning with the intention of getting future classification.
Data Available: Kaggle Repository.
License: Free Access.
Tools: Python, Jupyter, Sklearn, Pandas, Numpy, Matplotlib, Seaborn, Flask, PyQT5, Streamlit.
APIs: Flask (winepredict.azurewebsites.net/), Example for predict in API: (https://winepredict.azurewebsites.net/wine?tipo=1&fixed=7&volatile=1&citric=1&residual=45&chlorides=0.5&freesulfur=100&totalsulfur=200&density=1&ph=3.3&sulphates=1&alcohol=12)
Pricing: Free.
Access: Free Instance in Azure (Created in Feb2022) Time Limited. Without Athentication.
Test Data: Creation of test set, algorithm evaluation and cross validation
Evaluation: Perform the solution on the test data and present the classification precision metrics.
Availability: Github and Kaggle
Endpoints: Jupyter: https://github.com/esdfirls/Wine-Prediction/blob/main/Wine.ipynb , Flask API: https://github.com/esdfirls/Wine-Prediction/blob/main/app.py, Web App: https://github.com/esdfirls/Wine-Prediction/blob/main/streamtest.py, Jupyter Widget: https://github.com/esdfirls/Wine-Prediction/blob/main/wine_widget.py, Desktop App: https://github.com/esdfirls/Wine-Prediction/blob/main/Exec/winegui.py.
Flask API Json Result:
Jupyter Widget: http://localhost:8888/lab/tree/OneDrive/Data%20Science%20Knowledge/Projetinhos/Wine-Prediction/Wine.ipynb (at end)
Desktopp App Pyqt5: https://github.com/esdfirls/Wine-Prediction/blob/main/Exec/dist/winegui.exe
Web App in Streamlit: https://share.streamlit.io/esdfirls/wine-prediction/main/streamtest.py
