Projects
Saba Agizew Woldeamanuel
Smart Ad A/B Testing
A/B testing is a user experience research methodology. A/B tests consist of a randomized experiment with two variants, A and B., which are identical except for one variation that might affect a user's behavior. It includes the application of statistical hypothesis testing or "two-sample hypothesis testing" as used in the field of statistic
TELECOM USER DATA ANALYTICS
Exploratory Data Analysis (EDA) to get insights on the user activities on the telecom network. Assess user behavior, user experience, and user satisfaction on the network. And conclude whether the company is worth buying and figure out which areas to focus on to maximize user activity on the network.
PHARMACEUTICALS SALES PRICE PREDICTION
Applied machine learning models to forecast sales across multiple Rossman's pharmaceutical stores. The Data is time serious I used deep learning(LSTM) to predict future sales.
AMHARIC SPEECH-TO-TEXT ENGINE
Worked in a group of 7 to make an Automatic speech recognition system for the Amharic language. Layed out the MLOps pipeline using CML and DVC which had GPU runners integrated within Pull requests to allow easy training on AWS server.
Twitter Data Analysis
Analyses my Twitter data for sentiment and creates a Twitter sentiment analysis model. loads the data forking from 10xac/Twitter-Data-Analysis (github.com), extract and test it by using Travis CI, And build a dashboard by using MYSQL.
Breast cancer Causality Graph
worked in group 4. we Perform feature extraction by finding out the relationships of each variable with targeted variables. and the relationship between each other variables. And after that, we conduct the scaling and regression method.
Then we make a causal graph. To implement causal graphs we use CausalGraphicalModel is a python module for describing and manipulating Causal Graphical Models and Structural Causal Models.
Data engineering: speech to text data collection with kafka, Airflow and spark
worked in group 8 on speech to text data.The tool that we create should be deployed to process posting and receiving text and audio files from and into a data lake, apply transformation in a distributed manner, and load it into a warehouse in a suitable format to train a speech-to-text model.