Project status: Completed
The objective of this project is to develop a Natural Language Processing (NLP) model that provides cocktail recommendations based on user-provided ingredients. The model aims to leverage NLP techniques to analyze and understand the relationship between various cocktail ingredients and then make personalized drink recommendations.
- Natural Language Processing
- Feature Engineering
- Model Selection
- Data Visualization
- Data Wrangling
- Python
- Pandas
- Scikit-learn
- Natural Language Toolkit (NLTK)
- Streamlit
- Tableau
The dataset used in this project consists of a collection of cocktail recipes, including the ingredients used and instructions for preparation. The data was collected using the TheCocktailDb, a database that stores hundres of recipes for drinks and shares them with the usage of an API (a free version or a paid one). The dataset includes both alcoholic and non-alcoholic drinks, along with a variety of ingredients, clasiffiers (glass type, IBA, drink type, etc) and the measures for the ingredients.
You can access the API here: https://www.thecocktaildb.com/api.php?ref=apilist.fun.
- Data Collection: Gathering the cocktail recipes and associated data with the API
- Data Cleaning and Preprocessing: Removing duplicates, handling missing values, and preparing the data for NLP analysis
- Exploratory Data Analysis: Understanding the distribution of ingredients, common flavor profiles, and ingredient combinations
- Feature Engineering: Creating features based on ingredient similarity and drink categories
- NLP Model Development: Implementing word vectorization, similarity analysis, and recommendation logic
- Model Evaluation: Assessing the model's performance in providing relevant and accurate recommendations
- Data Visualization: Visualizing ingredient correlations, model performance, and user interactions
The NLP cocktail recommendation model successfully provides personalized drink recommendations based on user-input ingredients. The model utilizes NLP techniques to understand the relationships between ingredients and to offer tailored suggestions for cocktail recipes. The model's performance and user satisfaction have been positive, making it a valuable tool for cocktail enthusiasts and bartenders.
The recommendation model can be viewed here: https://final-project-brief-end-to-end-data-analytics-project-5ivblfah.streamlit.app/
LinkedIn: https://www.linkedin.com/in/axel-lopez-cabezas-213194170/
GitHub: https://github.com/redsun498