Wissenschaftliche Hilfskraft - Data4Sim Project
Universität Rostock - Collaboration with MotionMiners GmbH, Dortmund
01.2024 - 10.2024
What is Data4Sim project?
Developed methods to capture manual workflows using wearable sensors and convert them into simulation models. AI techniques, such as Convolutional Neural Networks (CNNs), identified employee activities, which were modeled with parameters like task durations and dependencies. These simulation models enable precise analysis and optimization of workflows, such as picking processes in logistics. They allow evaluation of changes in strategies, like picklist generation or product distribution, to improve efficiency and outcomes. Learn more
My Responsibilities and contribution:
- Video Annotation & Dataset Creation: Created a dataset by performing video annotation on 1.6 million frames, achieving an impressive accuracy of 95.89% in identifying and labeling key elements in the frames.
- Data Cleaning & Preprocessing: Reduced model training time by 20% through efficient data cleaning and preprocessing using Numpy and Pandas, optimizing data quality for machine learning models.
- Implemented Machine Learning Models: Developed and applied both supervised and unsupervised Hidden Markov Models (HMM) to analyze and predict patterns from the annotated data, improving model accuracy.
- Prediction Visualization: Visualized model predictions using Python libraries like Matplotlib and Seaborn to provide clear and insightful graphical representations of the results.
- Achieved High Precision: Successfully achieved a precision of 92.47% in model predictions, demonstrating the effectiveness of the implemented models and the quality of the dataset.
Working Student - Universität Rostock
10.2022 - 03.2023
Project Overview
The objective of this working student role is to manage the ETL process for geospatial data sourced from OpenStreetMap. The role involves developing a dynamic weight function for a graph representation of the road network, incorporating factors such as road segment lengths, traffic details, road types, and more. This dynamic weight function is then used to determine the shortest path within the road network effectively.
Contribution:
- Extracted OpenStreetMap (OSM) Data: Utilized the Python library OSMnx to extract detailed geospatial data from OpenStreetMap, enabling access to road networks, traffic data, and other relevant map features for further analysis
- Data Cleaning and Preprocessing: Applied Numpy and Pandas for cleaning and preprocessing raw geospatial data, handling missing values, correcting inconsistencies, and transforming the data into a usable format for analysis and model development.
- Developed Dynamic Weight Function: Designed and implemented a dynamic weight function that accounted for multiple factors such as traffic conditions, signal wait times, road types, and road segment lengths, enhancing the accuracy of pathfinding in the network.
- Traffic Analysis and Route Optimization: Performed in-depth traffic analysis to assess route efficiency, optimizing travel time and distance by calculating the shortest path based on real-time traffic data and dynamic weights, ensuring more effective and reliable route planning.
- Route Visualization: Used Python libraries such as Matplotlib and Folium to visualize the optimized routes, providing an intuitive graphical representation of the shortest paths on an interactive map to aid in better decision-making.
Junior Data Analyst - Loop Infosol, India
05.2020 - 03.2021
About LOOP INFOSOL
LOOP INFOSOL is a technology solutions provider specializing in eCommerce consulting, web and mobile development, UI/UX design, and digital marketing services. The company focuses on delivering client-centric, scalable, and innovative solutions to enhance user experiences and drive business growth.
My Key Contributions:
As a Junior Data Analyst, I contributed to the digital marketing services domain, assisting stakeholders of various automobile showroom franchises in making data-driven decisions. My role focused on delivering actionable insights to optimize sales and marketing strategies effectively.
Here are the tasks and responsibilities I handled during the job.
- Automated Data Extraction Processes: Developed and implemented SQL stored procedures to automate the extraction of automotive sales data, significantly improving data accuracy and efficiency while reducing manual effort.
- Sales and Customer Behavior Analysis: Leveraged advanced SQL queries to analyze sales trends and customer behavior, uncovering insights that informed strategic eCommerce and marketing decisions.
- Interactive Data Visualization: Designed and maintained dynamic Power BI dashboards for ad-hoc reporting, enabling stakeholders to access real-time, actionable insights to drive business growth.
- Collaboration for Data-Driven Decisions: Worked closely with senior sales and marketing teams to communicate data insights, empowering them to make informed decisions aligned with business objectives.
Bhavadip Dobariya
Data Enthusiast
© Copyright. All rights reserved.
Wir benötigen Ihre Zustimmung zum Laden der Übersetzungen
Wir nutzen einen Drittanbieter-Service, um den Inhalt der Website zu übersetzen, der möglicherweise Daten über Ihre Aktivitäten sammelt. Bitte überprüfen Sie die Details in der Datenschutzerklärung und akzeptieren Sie den Dienst, um die Übersetzungen zu sehen.