Brief Overview

EMM is UAE's mission to Mars launched in 2020. EMIRS is the spectrometer aboard EMM. We use the Level 3 calibrated data product from EMIRS to look for methane absorption in the spectra.  

Challenge

EMM is a particularly noisy dataset, as the orbit apoapsis is 43,000 km and periapsis is 20,000 km. Can we use the EMIRS Spectrometer data to search for the presence of Methane, and put an upper limit of methane detection?

Solution

Use De-noising techniques from established frameworks to reduce the noise in the data 

Use statistical clustering from the machine learning toolkit to search for absorption of methane at 1304 cm-1 in diurnal cleaned data from EMIRS

Establish detection/non-detection of methane by the EMIRS instrument oboard the Hope Mission. 

Set Upperlimit for Methane detection for the EMIRS instrument on EMM

Establish reliabe re-usable de-noising techniques for EMIRS Emissivity data. This framework can provide consistent, reliable data for future Machine Learning pipelines. 

GITHUB Repository

Research Paper in progress 

Topics: 

Data Ingestion, Data Sourcing, API, De-noising techniques, Outlier Detection, EDA, Statistical Analysis, k-means clustering, C-H Index, Linear Regression

Tools:

python, excel, EMM SDC api, TES data tool, pandas, matplotlib, seaborn, astropy, latex, overleaf 

Data:

Tes data is both structured and unstructured (csv, binary)

EMM data is structured (csv)