Please use this identifier to cite or link to this item: https://hdl.handle.net/10356/76142
Title: Data analytics on electricity consumption
Authors: Pan, Jiacong
Keywords: DRNTU::Science::Mathematics::Applied mathematics::Data visualization
DRNTU::Engineering::::Computer science and engineering
Issue Date: 2018
Abstract: Energy Research Institute @ NTU (ERI@N) has been collecting electricity meter readings for several buildings in Nanyang Technological University (NTU). This project aims to carry out data analytics on the electricity consumption in NTU to discover insights and anomalies in the data and replace anomalies in the data by estimates from the normal trend. The first part of the project was to visualize and analyze the data. Python libraries such as Pandas and Matplotlib were used to read and plot graphs of the data. With the visualizations, general trends such as weekly trends and daily trends were recognized. Anomalies which deviate from the trend were identified as well. The second part of the project was to use models to learn the normal trend and replace the anomaly data with estimates of the normal data. Firstly, the raw data was pre-processed to remove the anomalies and to obtain the training and test datasets. Secondly, two models, cubic spline and Long Short Term Memory network (LSTM), were configured to train with the training dataset. Lastly, the trained models were used to predict the test set. Based on the actual test set and predicted results, evaluation metrics such as Root Mean Squared Error and Mean Absolute Error were calculated and the performance of the models were discussed.
URI: http://hdl.handle.net/10356/76142
Rights: Nanyang Technological University
Fulltext Permission: restricted
Fulltext Availability: With Fulltext
Appears in Collections:SCSE Student Reports (FYP/IA/PA/PI)

Files in This Item:
File Description SizeFormat 
Pan Jiacong FYP Final Report.pdf
  Restricted Access
1.43 MBAdobe PDFView/Open

Page view(s) 50

65
checked on Oct 29, 2020

Download(s) 50

34
checked on Oct 29, 2020

Google ScholarTM

Check

Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.