University of Texas at Dallas, Texas, United States
Abstract: The climate crisis is the greatest challenge facing humanity. About 20 million people a year are forced from their homes due to climate change, air pollution alone is estimated to cause over 4 million premature deaths a year worldwide, 785 million people lack access to safe and clean drinking water. There is an increase in demand for understanding the environmental factors including air quality index, water quality, UV index, etc. There is a growing need to quantify these parameters. The purpose of this innovation is to create a quantitative metrics to better assess environmental health which can adversely impact human health directly or indirectly. One Health metric aims to give a consolidated quantitative measure of the conditions of a particular location taking into consideration the air quality, water quality and the soil health. Soil health metric provides the real time information of the soil to the farmers which allows them to take the appropriate decisions for their land and optimize the yield from the land while the soil is not hampered with.
The geospatial data, data from government organizations and educational institutions are used for the training and testing of the models. The amounts chloramines, sulphates, pH, organic carbon, and other solids present in the water are used as the input parameters for determining whether the water sample is potable or not. The air that we breathe too has toxic compounds like nitrous oxide, carbon monoxide, ozone, etc. Using the quantities of these compounds in the atmosphere an aggregate air quality index is generated and this dataset is used as a training dataset for the model to predict the index a particular location by taking the location as an input. The third aspect, i.e, soil health is also quantified in a similar approach, by measuring the nitrates, bulk density, sulphates, moisture and using this data to train the model and generate an index for the soil. Combining the three indices, a consolidated index is obtained. Based on the consolidated index, the location is graphically represented in a red (danger) to green (safe) zone. Decision tree (accuracy = 0.60), KNN (accuracy = 0.62), SVM (accuracy = 0.70), random forest (accuracy = 0.68) models were used to predict water quality index. For air quality index, different regression models are explored and RMSE and R-squared values are used as the success metrics to determine the suitable model to be used. We’re building and integrating our group’s sensor data into this model and getting an overall climate and environmental health index.