Appliance Energy Prediction

Capstone project for Udacity Machine Learning Nanodegree

Early Work

• Research Paper
  
• GitHub repo

Dataset Link

• Appliance Energy Data

Dataset Information

• Number of instances: 19,735
• Number of attributes: 29

Attribute Information

1. date: year-month-day hour:minute:second
2. T1: Temperature in kitchen area, in Celsius
3. RH_1: Humidity in kitchen area, in %
4. T2: Temperature in living room area, in Celsius
5. RH_2: Humidity in living room area, in %
6. T3: Temperature in laundry room area
7. RH_3: Humidity in laundry room area, in %
8. T4: Temperature in office room, in Celsius
9. RH_4: Humidity in office room, in %
10. T5: Temperature in bathroom, in Celsius
11. RH_5: Humidity in bathroom, in %
12. T6: Temperature outside the building (north side), in Celsius
13. RH_6: Humidity outside the building (north side), in %
14. T7: Temperature in ironing room, in Celsius
15. RH_7: Humidity in ironing room, in %
16. T8: Temperature in teenager room 2, in Celsius
17. RH_8: Humidity in teenager room 2, in %
18. T9: Temperature in parents’ room, in Celsius
19. RH_9: Humidity in parents’ room, in %
20. T_out: Temperature outside (from Chievres weather station), in Celsius
21. Pressure: (from Chievres weather station), in mm Hg
22. RH_out: Humidity outside (from Chievres weather station), in %
23. Wind speed: (from Chievres weather station), in m/s
24. Visibility: (from Chievres weather station), in km
25. T_dewpoint: (from Chievres weather station), °C
26. rv1: Random variable 1, non-dimensional
27. rv2: Random variable 2, non-dimensional
28. Lights: energy use of light fixtures in the house in Wh
29. Appliances: energy use in Wh (Target Variable)

Software Requirements

Python version: 3.5

The software requirements are mentioned in the requirements.txt file in top level directory. To install these dependencies, use the command:-

pip3 install -r requirements.txt

The installation of SciPy would fail on Windows because of absence of Numpy + MKL. To install dependencies on windows, follows the steps:-

1. Download NumPy + MKL for the mentioned python version from here.
2. Remove the Numpy's entry from requirements.txt.
3. Run the command: pip3 install -r requirements.txt.

Running the software

To run the software, navigate to the top level directory and type

jupyter notebook predict_energy_consumption.ipynb