Climatic, population and daily water consumption dataset of sub tropical humid region of Pakistan
Seven years of water consumption data, along with population data, were manually collected in collaboration with the local municipality office. This data was then combined with climatic data to model the proposed machine learning algorithm. The weather data was recorded for a period of 7 years using precise meteorological instruments installed in Islamabad at coordinates 33.64° N and 72.98° E, with an elevation of 500 meters above sea level.
The novel training dataset contains seven variables, providing daily information on population count, water consumption in cubic meters, ambient temperature, dewpoint temperature, percentage humidity, wind speed, and precipitation in inches, spanning seven years. This dataset offers valuable insights into the local population of selected sites and their water consumption, information that was not previously curated.
The population data is categorized into two sub-groups: permanent residents and seasonal residents. Permanent residents reside in residential areas throughout the year, while seasonal residents live in the area for 8-10 months each year.
The urban site selected for water consumption analysis in this dataset comprises three overhead tanks (OHTs) filled by nine different water pumps installed at distant locations.
The data collection was carried out using the following instruments:
1. Campbell CS 215l: Used to record ambient temperature and relative humidity.
2. NRG 40H Anemometer: Utilized for recording wind speed.
3. Campbell CS 100: Employed for measuring barometric pressure.
Overall, this dataset and the information gathered from the instruments will be instrumental in developing and improving the proposed machine learning algorithm for water consumption prediction and analysis.
The dataset is available in both Excel and CSV formats, allowing it to be analyzed and utilized for training algorithms with any machine learning tool. We trained a machine learning model using this dataset in the Python programming language.