Contributors:
- Mehwish Nabi
- Yaou Hu
- Nate Puangpanbut
A project to predict the heating load of various types of buildings and to investigate the role of features in heating load prediction. This project is a group study according to DSCI 522 (Data Science workflows); a course in the Master of Data Science program at the University of British Columbia, Fall 2022.
Building towers or any building structure nowadays is not difficult if you can afford it, but building it to be the most memorable and efficient is another story. When considering building new towers or skyscraper buildings, it will be great if we know exactly what building parameters relate to their energy efficiency. As a result, we would be able to design not only a magnificent building to remember but also a renowned energy-efficient building.
In this project, we gather simulating data of building heating load versus multiple building features created by Angeliki Xifara ([email protected], Civil/Structural Engineer) and was processed by Athanasios Tsanas ([email protected], Oxford Centre for Industrial and Applied Mathematics, University of Oxford, UK), and is made available under the UCI Machine Learning Repository and can be found here specifically this file.
The data set contains 12 different building shapes simulated in Ecotect. The buildings differ in the glazing area, the glazing area distribution, and the orientation, amongst other parameters. Various settings were simulated as functions of the aforementioned characteristics to obtain 768 building shapes. The data set comprises 768 samples and eight features.
Specifically:
X1 Relative Compactness
X2 Surface Area
X3 Wall Area
X4 Roof Area
X5 Overall Height
X6 Orientation
X7 Glazing Area
X8 Glazing Area Distribution
y1 Heating Load
y2 Cooling Load
Given building-related features such as Relative Compactness', 'Surface Area', 'Wall Area', 'Roof Area', 'Overall Height', 'Orientation', 'Glazing Area', and 'Glazing Area Distribution', how accurately can we predict the 'Heating Load' of the building? And to check contribution level of each feature to the 'Heating Load' of the building.
We split the data set into train and test data sets with 70:30 ratio. An EDA is performed to reveal the distribution and relationship among the features to the response. The result of EDA and its correlation coefficient map is analyzed to reduce some unnecessary features. The modeling fitting is performed with various types of models such as
- KNN
- Ridge
- Decision Tree
- Support Vector Machine
- Random Forest
- XG Boost
Each model is fitted with the same transformation and pipeline and is reported as the mean and standard deviation of the cross-validation score. The best cross-validation score model is selected to be the final model to perform prediction on test data to see how it fits and performs. Finally, all the models are saved as pickled files.
Partition the data set into training and test sub-data sets
The data set was divided into train and test sets, with 70% train data and 30% test data.
Exploratory data analysis with the train set
The exploratory data analysis was conducted through the following steps:
-
Load the necessary packages and split the data into train and test sets. NaN data were dropped;
-
EDA was performed on the train set. We checked the data types and found no missing values. Then we checked the data distribution through bar plots, value_counts, correlations, and pairwise scatter plots. Through the EDA, we could identify that all the variables are numeric. We kept all the features for subsequent analysis.
-
Based on the above analysis, we aim to do a supervised machine learning model with ‘Heating Load’ as the target.
The results of the EDA can be found here.
The final report can be found here.
There are two suggested ways to run this analysis:
1. Using Docker
note - the instructions in this section also depends on running this in a unix shell (e.g., terminal or Git Bash)
To reset the repo to a clean state, with no intermediate or results files, run the following command at the command line/terminal from the root directory of this project:
docker run --rm --platform linux/amd64 -v $PWD:/home suraporn/energy_efficiency_analysis:v0.46.0 make -C '/home' clean
To replicate the analysis, install Docker. Then clone this GitHub repository and run the following command at the command line/terminal from the root directory of this project:
docker run --rm --platform linux/amd64 -v $PWD:/home suraporn/energy_efficiency_analysis:v0.46.0 make -C '/home' all
Please be noted that,
- You may remove
--rm --platform linux/amd64, if you are not using Mac M1 architecture. - You may put some specific path to your mount and rin the docker image and container such as
$ docker run --rm -v /$(pwd):/C/Users/yaouh/energy_efficiency_analysis suraporn/energy_efficiency_analysis:v0.46.0 make -C C/Users/yaouh/energy_efficiency_analysis clean
$ docker run --rm -v /$(pwd):/C/Users/yaouh/energy_efficiency_analysis suraporn/energy_efficiency_analysis:v0.46.0 make -C C/Users/yaouh/energy_efficiency_analysis all
2. Without using Docker
To replicate the analysis, clone this GitHub repository to your local:
git clone https://github.com/UBC-MDS/energy_efficiency_analysis.git
Navigate to your local repository and prompt the command line and run:
conda env create --file energy_project.yaml
The new environment energy_project will be created in your conda environment, and we will use this as the main environment to run the analysis.
Activate the new environment by:
conda activate energy_project
To reset the repo to a clean state, with no intermediate or results files, run the following command at the command line/terminal from the root directory of this project:
make clean
To replicate all of the analysis, run the following command at the command line/terminal from the root directory of this project:
make all
The Python and Python packages can be found here.
MIT License Copyright (c) 2022 Mehwish Nabi, Yaou Hu, Nate Puangpanbut
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation, the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the software, and to allow those to whom the software is provided to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the software.
THE SOFTWARE IS PROVIDED "AS IS," WITHOUT ANY EXPRESS OR IMPLIED WARRANTY.INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, AND NON-INFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES, OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF, OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE
[1] Harris, Charles R., K. Jarrod Millman, Stéfan J. van der Walt, Ralf Gommers, Pauli Virtanen, David Cournapeau, Eric Wieser, et al. 2020. “Array Programming with NumPy.” Nature 585 (7825): 357–62. https://doi.org/10.1038/s41586-020-2649-2.
[2] Pedregosa, F., G. Varoquaux, A. Gramfort, V. Michel, B. Thirion, O. Grisel, M. Blondel, et al. 2011. “Scikit-Learn: Machine Learning in Python.” Journal of Machine Learning Research 12: 2825–30.
[3] The pandas development. 2020. Pandas-Dev/Pandas: Pandas (version latest). Zenodo. https://doi.org/10.5281/zenodo.3509134.
[4] Tsanas, Athanasios, and Angeliki Xifara. 2012. “Accurate Quantitative Estimation of Energy Performance of Residential Buildings Using Statistical Machine Learning Tools.” Energy and Buildings 49: 560–67.
[5] VanderPlas, Jacob, Brian Granger, Jeffrey Heer, Dominik Moritz, Kanit Wongsuphasawat, Arvind Satyanarayan, Eitan Lees, Ilia Timofeev, Ben Welsh, and Scott Sievert. 2018. “Altair: Interactive Statistical Visualizations for Python.” Journal of Open Source Software 3 (32): 1057.