Flora Genie - A Personalized Plant Recommendation System

Flora Genie is a personalized plant recommendation system designed to assist amateur gardening enthusiasts in selecting the most suitable plants for their home or garden.

By specifying criteria such as plant category, light and water requirements, plant form, foliage description, and more, users can receive tailored recommendations to suit their needs.

Features

• Personalized plant recommendations based on user-specified criteria.
• Comprehensive plant database including textual data and images.
• Multiple clustering algorithms for effective plant grouping.
• Similarity metrics for accurate recommendation generation.

File Structure

Flora-Genie/
├── Algorithms and Scripts/ 
│   └── ...
├── Application/
│   ├── app.py
│   └── ...
├── Data Files/
└── requirements.txt

Algorithms and Scripts: Contains Python scripts for the core functionality of the project such as clustering, recommendations, data preprocessing, data extraction, and utility functions.

Application: Contains files related to the web application.

Data Files: Contains the extracted plant dataset and related files.

requirements.txt: File listing all the necessary Python packages for the project

Data Sources

India Flora Online

India Plants

IMPPAT: Indian Medicinal Plants, Phytochemistry And Therapeutics

Algorithms and Methodologies

Pipeline

Data Extraction

• Compiled plant data using libraries like BeautifulSoup, Selenium, and urllib from three different data sources.

Data Preprocessing

• Transformed categorical columns into one-hot encoding.
• Applied TF-IDF for feature extraction from textual data.
• Performed tokenization, stop word removal, and lemmatization.

Clustering Algorithms

• KMeans and KMeans++: Identified optimal clusters (36) with a silhouette score of 0.2095.
• Spectral Clustering: Showed imbalance with most points in one cluster.
• Hierarchical Clustering: Observed distinct clusters from the dendrogram.

Similarity Metrics

• Cosine Similarity
• Jaccard Similarity

Recommendation Generation

• Used collaborative filtering and content-based similarity with TF-IDF for personalized recommendations.

Results

• Balanced clustering distribution using KMeans and KMeans++ algorithms.
• Accurate personalized plant recommendations based on user criteria.
• Visual insights provided through plant images enhance user experience.

Run Locally

Clone the repository

git clone https://github.com/kumaranjalij/Flora-Genie.git

Go to the project application directory

  cd Flora-Genie/Application

Install dependencies

  pip install -r requirements.txt

Run the flask application

  python -m flask run

Screenshots