Light the Path

The COVID-19 pandemic initiated changes in human population movements and activities around the world. Your challenge is to use Earth observations to explore how human activity and regional land-based human movement patterns may have shifted in response to COVID-19.

PILOT LIGHT

Summary

Our project, Pilot Light, is based on the satellite VIIRS/BlackMarble and IASO data. We took night time images from March 20th 2020 to May 5th 2020. The night time images gave us amazing data directly related to the spread of COVID-19 cases. We estimated the fluctuation in human activity during that period in the Eastern US by observing the changes in light intensity of the images over certain period of time.Now looking at the COVID-19 cases in that region but 2 weeks ahead (in order to let the disease incubate) of time period of the night time images gives us some amazing relation between light intensity of night time images and COVID-19 cases.

How We Addressed This Challenge

CHALLENGE : LIGHT THE PATH 

Pilot Light helps in understanding population movement and it's relation with  the COVID-19 pandemic. By integrating Machine learning with COVID-19 cases we can predict the increase and decrease in disease cases by directly comparing it with the current night time images. This project can be made more accessible by showing the data for more cities rather than the few cities we tried during these 48 hours (Prototype). This data can eventually be helpful for government in order to implement rules and preventing future outbreaks. Our project is just the start in order to become a website/application which can save many lives.


How We Developed This Project

We saw an article (mentioned in the presentation) which compared the night time images of  Wuhan, China before and after COVID-19. The difference in the images was shocking and really intriguing. This led us to take advantage of the night-time image datasets (or other data which is helpful in predicting human activities ie: CO emission) in order to provide an open source way to make that data accessible and readable for everyone. We used Javascript to make a website and Neustar for the domain. We used QGIS, ENVI for rendering the image (which were in HDF file format). To retrieve the light intensity from the images we used FIJI software. We used Google Collab for Machine Learning (incomplete). We faced a lot of problems in retrieving up to date night images.  Google Earth engine had datasets only till January even though it was claimed to be up to date. This was true for many data we found and some data would not be available until early June. On the other hand, we were amazed to see how our pitch idea was correct when we saw the graph between light intensity of Night time images and COVID-19 cases (2 weeks ahead of the time of night light images in order to give type for developing symptoms). This was our biggest achievement, seeing everything coming together and giving the result what we expected.

Our project PILOT-LIGHT website : Detailed presentation

https://covid19.spaceappschallenge.org/challenges/covid-challenges/light-path/teams/pilot-light/stream

(Mentioned in the updates)

Team Presentation

Demonstration By Our Team Pilot Light

Nikunj Aggarwal

James Schaefer-Pham

Guillaume Landry

Aniket Mishra

Tags
#Nighttime #BlackMarble #COemission #VIIRS #Webapp #DataAnalysis
Global Judging
This project was submitted for consideration during the Space Apps Global Judging process.