Alejandro M. Torres Castellanos

Knight Foundation School of Computing and Information Sciences

Lecture Information:
  • March 10, 2021
  • 4:00 PM
  • Zoom

Speaker Bio

Alejandro M. Torres Castellanos is a Master’s student at the Knight Foundation School of Computing and Information Sciences (KFSCIS) and a member of CREST CAChE, Florida International University (FIU). He obtained his Bachelor’s degree and Master’s degree in Civil Engineering at FIU. His subject of interest is Artificial Intelligence, Machine Learning, Deep Learning, and Data Science.


Water quality is a very active subject of research in the water science field, where its importance includes maintaining the environment, managing wastewater, and securing fresh water. However, due to the increase of human developments, it has led to problems that are affecting the ecosystem. Such issues can be seen in the Florida Keys, where coast waters suffer from events like the red tide. Motivated by these problems, this research aims to a solution for understanding the coastal water of the Florida Keys. By making use of a large dataset provided by FIU South East Environmental Research Center, where 25 years of data were cleaned and structured.

This research used machine learning methods to find a correlation between water quality dataset and profile measurements dataset. To achieve this objective, this research first went through structuring a readable dataset of the profile measurements that could be used in the analysis. To obtain this dataset, a Python code was developed to clean, rescue, and structure the data. Once completed the dataset, coming from the profile measurements, the next step was to find the correlation. To get a correlation between two datasets, this study proposes the use of regression coefficients coming from four different measurements in the profile dataset. Having the coefficients, they were clustered using K-means and compared with the clustered result of the chemical dataset. However, since K-means depends on optimal cluster values, different optimal cluster analyses were completed such as the elbow method, gap statistic, and silhouette analysis. After obtaining the optimal cluster value, then an independency test was carried out on datasets. It was found that using regression coefficients was not the right approach for finding a correlation between the profile and chemical dataset. Lastly, this study also built a water drone in the form of an airboat, which can collect data and can be controlled through an android app.