Legendnic - Space Apps Challenge

Legendnic| Spot That Fire V3.0

Spot That Fire V3.0

Recent wildfires worldwide have demonstrated the importance of rapid wildfire detection, mitigation, and community impact assessment analysis. Your challenge is to develop and/or augment an existing application to detect, predict, and assess the economic impacts from actual or potential wildfires by leveraging high-frequency data from a new generation of geostationary satellites, data from polar-orbiting environmental satellites, and other open-source datasets.

Spot The Fire API

Summary

For the base system, it is providing an API for user to do ETL on dirty datasets or retrieve the processed satellite datasets provided by NASA. User can used the cleaned datasets for machine learning training, or cleaning up the raw datasets available in the server. Since it was built on the top of API, therefore any users can used the provided API to create their own interface. Futhermore, it can be expanded in terms of data prediction and detection, built-in machine learning training and testing, and providing high level data visualization for important decision making.

How I Addressed This Challenge

This program is a stand-alone API project. Therefore, it does not related or constraints with specific interface. By this way, many user can customize their own interface only just using this API.

In a sense, ETL process is tiring process and time consuming. In this API, the ETL process being done automatically and uniform, despite the satellite data are variables. Not only users can avoid repetitive and never-ending tasks, users can focus on algorithmic construction; or directly making data visualization for important decision. In the future, if this system would be complete, it will boost up the development of algorithms, without truly consuming time for directly configure algorithms in lower level perspective

How I Developed This Project

Currently pursuing master degree in computer science, specifically in deep learning and computer vision. However, my ability to play around with dirty and tedious datasets other than images is quite lower than expected. Therefore, this challenge is really fitting for me since it use time series, images, and meta data to actually produce real world impact. I am using python language and libraries only, since Python have mass amount of data science tools, such as geopandas, pytorch, and matplotlib. Furthermore, python is well developed language for easy to integrate with any kind of devices, such as desktop and mobiles, either by using web based system or just by providing API. I am using high performance desktop for scientific purpose.

There are challenges, however. FIrst of all, I have hard time to understand the data provided. It takes longer time than it supposed to be. Despite using high performance computer, my computer still not supporting larger datasets, either for calculation, extraction, or visualization.

Still, I have my own achievements, such as involving with harder and dirty datasets. Thus, it forces me to elevate my critical thinking in solving the problem. Since I am doing everything by my own, I guessed I learn more in developing API, which is extensible later for my own purpose. The space app challenge become my benchmark in terms of evaluating my problem solving skill.

How I Used Space Agency Data in This Project

I am using FIRMS database, which providing the latitude and longitude. by this way, I am reprojecting the coordinates using Python Basemap. Futhermore, my system can chunking the large data into several categories, depending on what is being desired. From there, it can be stacked together for machine learning training. Here is my project https://github.com/Legendnic/spot_the_fire