ManhTin / plant-recommender-datasource

Data Analysis and Dataset generation for the plant recommender backend

Plant Recommender App

Datasourcing for the Homeplant Recommender System project in Data Integration 22/23 at WWU Münster

Data Sources

• https://plants.usda.gov/home for outdoor plants
• https://howmanyplants.com/ for indoor plants
• https://nobleapps.noble.org/plantimagegallery/ for pictures of outdoor plants

Cleaned Data

The cleaned dataset contains 1682 indoor and outdoor plants with following attributes:

Attribute	Indoor	Outdoor
active_growth_period	-	x
bloom_period	-	x
climate	x	-
common_name	x	x
difficulty	x	-
drought_tolerance	x	x
duration	-	x
family	-	x
family_common_name	-	x
flower_color	-	x
foliage_color	x	x
foliage_porosity_summer	-	x
foliage_porosity_winter	-	x
frost_free_days	-	x
fruit_color	-	x
growth_habit	x	x
growth_rate	-	x
height	x	x
humidity	x	-
image	x	~
leaf_shape	x	x
lifespan	-	x
light	x	-
origin	x	x
ph_minimum	-	x
ph_maximum	-	x
scientific_name	x	x
temperature	x	-
toxicity	x	x
type	x	x
width	x	-

Installation and Setup

Getting started

Prerequisites

• Python 3.10
  • We recommend to use miniforge (or miniconda). Since you can manage Python versions and virtual environments with it. Follow the setup steps below if you use miniforge (or miniconda)
  • after successful installation of miniforge (or miniconda) the command conda should be available
  • follow the setup guide from step 4 if you use another method of managing virtual environments
• Setup project
  1. cd into this directory
  2. Create virtual environment for project: conda create --name plant-recommender-datasource python=3.10
  3. Activate virtual environment of project: conda activate plant-recommender-datasource
  4. Install dependencies into virtual env: pip install -r requirements.txt
  5. Start jupyter server: jupyter notebook
• Run scraper
  1. howmanyplants.com scraper: python scrapers/how_many_plants_scraper.py -> exports to data/...