Updated docs

  - matplotlib

  - xlrd

  - jupyter

  - jupyter_contrib_nbextensions

  - tqdm

  - scikit-learn

  - pip

---

title: Datasets and system description

order: 5

hasequations: true

title: v1 Dataset and system description

order: 1

hasequations: false

---

This is the data set for *Chiller 1* at the Engineering Science Building. Parameter readings were recorded from January 1, 2018 through December 31, 2018. Measurements were recorded at a 5 minute interval.

The preprocessed dataset is available [here][4].

## Data extraction

The data are made available through the MetaSys application. The application is responsible for aggregating sensor readings in Engineering Science Building. It allows upto 14 days' of data and 10 fields to be viewed and copy-pasted to a spreadsheet at once.

[1]: https://en.wikipedia.org/wiki/Arden_Buck_equation

[2]: https://en.wikipedia.org/wiki/Dew_point#Calculating_the_dew_point

[3]: https://journals.ametsoc.org/doi/pdf/10.1175/JAMC-D-11-0143.1

[4]: https://vanderbilt.box.com/s/s85evxq7gk9mq43i74tjhrlormki18tx

---

title: v2 Dataset and system description

order: 1

hasequations: false

---

The v2 dataset is created by downloading trends created in the BuildingLogiX Data eXchange (BDX) [hosted at Vanderbilt][1]. Trends are a collection of measured variables in a single table. In particular, the following trends are used for making the models:

```

1. 2421 Chiller 1

2. 2623 Chiller 2

3. [2621 Cooling Tower 1][6]

4. [2622 Cooling Tower 2][5]

5. [2422 ESB HVAC Control (Chiller 1)][4]

6. 2841 ESB HVAC Control 2 (Chiller 2)

```

Trends can be downloaded using the [`bdx`][2] python package.

The naming conforms to the same conventions as in [`v1` dataset][3]. No preprocessing is done to convert to SI units.

[1]: https://facilities.app.vanderbilt.edu/trendview

[2]: https://git.isis.vanderbilt.edu/ahmedi/bdx

[3]: ../v1/dataset.md

[4]: https://vanderbilt.box.com/s/0xm8hvtyx9cwtclgbe6265jp941ll2tr

[5]: https://vanderbilt.box.com/s/dtnsr9919wre921ko9wafq8bargmskf2

[6]: https://vanderbilt.box.com/s/o1hqrq9iwcuknt2vieq541ez7099t6i6

 No newline at end of file

Requires python 3.7. However, python 3.5+ shoould work fine. To install dependencies:

```bash

> pip install -r requirements.txt

# or

> conda install --file requirements.txt

conda env create -f dev.yml  # for development

```

One requirement is `IPyVolume` for 3D plots. See installation instructions [here][2].

Saving GIFs using `IPyVolume` requires [ImageMagick][1] with legacy options (i.e. the `convert.exe` command) enabled.

Currently `IPyVolume` does not work with Jupyter Lab. Instead use Jupyter Notebook to view those plots.

# Source Organization

* `ioops`: contains functions to convert data sources into `csv` format for use by `pandas`. Call `python -m ioops XLSX_FILE` to convert the excel file to `csv`.

## Python packages

* `preprocess`: contains functions to clean up csv data for analysis. Call `python -m preprocess` to process all `csv` files in `../SystemInfo`.

* `preprocessing`: contains functions to clean up csv data for analysis. `v1` is for the older version where excel files were provided. `v2` is for data downloaded from BuildingLogix Data Exchange. Currently it is used as-is. To use, call `python -m preprocessing.v1.to_csv` and `python -m preprocessing.v1.cleanup` on Excel files.

* `thermo`: contains conversion functions for thermodynamical quantities (temperature, pressure etc.).

 No newline at end of file

* `utils`: Data wrangling and convenience functions.

* `systems`: Contains definitions of various reinforcement-learning environments that conform to the OpenAI gym interface.

* `controller`: The script used to control ESB condenser water temperature. Main entry point for production usage. See Repository README for more details.

## Notebooks

* `baseline_condenser`: Using a feedback controller without machine learning to test control of HVAC systems.

* `Models-v2`: Generating data-driven models from the [`v2` dataset][2], which has been preprocessed. See `docs/datasets/v1/` for more information.

* `RL-Cooling Tower` and `RL-Condenser`: Formulation of RL environments using `v2` datasets.

* `Models-v1`: Generating data-driven models from the [`v1` dataset][1], which has been preprocessed. See `docs/datasets/v1/` for more information.

* `Relationships`: Looking at various statistical metrics between fields in `v1` datasets.

* `Trends`: Plotting time series in `v1` datasets.

[1]: ./docs/datasets/v1/dataset.md

[1]: ./docs/datasets/v2/dataset.md

 No newline at end of file