Space Physicist & Data Scientist

Formal Education

• Ph.D. in Space physics, Bahir Dar University, Ethiopia (February 2021)
• M.Sc. in Space physics, Bahir Dar University, Ethiopia (June 2011)
• B.Sc. in Physics, Bahir Dar University, Ethiopia (June 2009)

Data Science Education

Google / Advanced Data Analytics Professional Certificate

March 2024 - May 2024, Online (Coursera)

An extensive Three-month job-ready Google Career Certificate training. Demonstrated hands-on experience with advanced Data Analytics practices and tools, including data science workflows, Python programming and Tableau. Confidence in exploratory data analysis; statistical data analysis; regression analysis and data modeling, and machine learning basics.

Python for Data Science and Machine Learning

January 2024 - February 2024, Online (UDemy)

The Complete Python Bootcamp From Zero to Hero in Python

October 2023 - December 2023, Online (UDemy)

Softwares and Skills

• Programming Languages: MATLAB, Python
• Python Packages: numpy, Pandas, Scipy, seaborn, Matplotlib, statsmodels, scikit-learn
• Machine Learning Models: regression (linear, logistic), Naive Bayes, decision trees, random forest, AdaBoost, XGBoost
• Data Science (Data cleaning, Data structuring, Data analysis)

Work Experiance

• Postdoctoral Researcher, Czech Academy of Sciences, Czech Republic, April 2023 – present.
• Assistant Professor of Space physics, Debre Tabor University, Ethiopia, January 2021 – March 2023.
• Lecturer at Debre Tabor University, Debre Tabor, Ethiopia, July 2011 – December, 2016.

Data Analytics Projects

• Classification of Waze data: Built decision tree, random forest, and XGBoost to predict Waze user churn

Waze App User churn project

About this project

• The goal of this project was to use Python to build a dataframe, perform a preliminary inspection of the dataset, and inform the other team members of the results.
• At this stage, we were focused on inspecting user data to uncover any important relationships in the driving data for retained and churned users.
• This also involved creating a pandas dataframe for data learning, future EDA, and statistical findings.
• The insights from this project will help guide the next steps for the analytics department.

Key Insights

• This dataset contains approximately 82% retained users and 18% churned users.
• The label column is missing 700 values.
• Retained users used the app over twice as many days as churned users in the last month.
• Churned users had more drives overall in fewer days, and their trips were longer in both distance and duration.
• The users in this dataset drove a lot of mileage overall. This may be worth exploring more to determine a more complete user profile.

Space Science Projects

• Upper atmosphere dynamics
• Coupling between Ionosphere and atmosphere
• Ionospheric irregularities
• Satellite data processing and analysis

Main outcomes (Space Physics)

Modeling Plasma drift velocity

Publication

To understand the dynamics of the equatorial ionosphere and mitigate its effect on radio wave propagation, vertical ion drift velocity empirical models have been developed using limited ground‐based and/or space‐based observations. These models, however, have not yet been validated in detail using recent observations for sufficiently different longitudinal sectors. In this paper we have evaluated the performance of two empirical models and also propose a simplified vertical drift velocity model based on basic physics laws (i.e., Ampere’s and Ohm’s laws) that we call it parameterized drift velocity (PDV) model. These models have been applied to estimate the Eregion electric field and the associated F region E×B drift velocity using observed horizontal magnetic fields, due to equatorial electrojet current, as model driver input. Drift velocities obtained from these models are compared with the Communication/Navigation Outage Forecasting System (C/NOFS) satellite in situ vertical drift observations for different longitudinal sectors.

Valdation of International Reference Ionosphere Model (IRI-2020 TEC)

Publication

We evaluated and reported the International Reference Ionosphere (IRI) model’s vertical Total Electron Content (TEC) regional profile. Diurnal, monthly, seasonal, and storm-time characteristics of IRI estimates over the equatorial region’s ionosphere are validated. We compared the vertical TEC derived from IRI-2020 and its predecessor, IRI-2016, with the GPS-TEC measurements.

Ionosphere-Thermosphere Response to the January 14, 2022 Moderate Storm: ICON Satellite Observations

We present a storm time Ionosphere-Thermosphere profile to the January 14, 2022 Moderate Storm under Solar Minimum condition. ICON satellite observations (ion density and neutral wind speed) are used for the storm time analysis. The study has been conducted using ionosphere-thermosphere parameters from a bout 85-600 km altitude which allows to look storm responses of different layers of ionosphere unlike the most common single-layer analysis.