• Flask
  • pip install flask
• Flask-SQLAlchemy
  • pip install flask-sqlalchemy
• PostgreSQL server
  • Installation is OS dependant
• PostgreSQL-server-dev-x.x
  • Installation is OS dependant sudo apt-get install postgresql-server-dev-9.1 for Ubuntu/Mint
• pyowm
  • pip install pyowm
• configparser
  • Probably installed by default
  • pip install configparser
• psycopg2
  • pip install psycopg2
• scikit-learn
  • pip install -U scikit-learn

1. Clone the repo onto your computer:

• git clone https://github.com/cruncher19/Ubicomp-GPP

2. Create a user and database in postgres for the prediction system to use

• in my case I used user: greenhouse, database: greenhousePower
• don't forget to give your user permissions on your new database

3. Change the db_uri: db_uri: postgresql://greenhouse@localhost/greenhousePower setting in the config file to point to your postgres instance

• should be of the form: db_uri: postgresql://<username>@localhost/<database>

4. Run python initdb.py to initialize the postgres database
5. Call python routes.py to start the server
6. You're good to go!

You can store power information in the database by POSTing to the server like this: http://localhost:5000/storePowerProduction?powerLevel=50 You can turn off all devices connected to the power relay arduino like this: http://localhost:5000/turnOffAllDevices You can turn on all devices connected to the power relay arduino like this: http://localhost:5000/turnOnAllDevices

run python predictor.py in the repository's directory.

##Solar power website scraper usage python PowerScraper.py

The above command will print the time the site was last updated, the current daily power production and add the power information to the database via the REST API.

This currently uses a fake dataset to build a regression model and then uses the current weather conditions with the model to estimate power production.