umitkaanusta
Funding Links: https://github.com/sponsors/umitkaanusta
- Name: Ümit Kaan Usta
- Location: Istanbul
- Company: @amzn Amazon
- Kind: user
- Followers: 113
- Following: 264
- Total stars: 427
- Repositories count: 19
- Created at: 2022-11-15T16:54:28.871Z
- Updated at: 2024-05-01T10:09:36.883Z
- Last synced at: 2024-05-01T10:09:35.455Z
GitHub Sponsors Profile
Intro
Hi, I'm Umit from Istanbul! Currently interning at Amazon where I build models to forecast, plan, and understand our supply chain - and a rising senior CS student at Marmara Uni.
Twitter
LinkedIn
What I build
I build tools to understand human behavior with data science, specifically focusing on social networks, information diffusion, political information operations, and conflicts.
I'll also start contributing to Graph Theory / Social Network Analysis libraries with implementations of different algorithms from academic papers.
My goals
As a student, I will use the funds to get the needed data and computing power, and to secure time to contribute to Open Source. Particular goals are:
[ ] 3 Contributions to top Graph libraries (e.g. NetworkX) with algorithms from academic papers
[ ] Performance updates to reddit-detective
[ ] Campaigning to increase adoption for reddit-detective, the focus is on getting some news outlets to use it to track US Elections discussions and disinformation campaigns.
[ ] A project offering explainability to time-series models
[ ] A project modeling casualties in modern warfare to help researchers and journalists have a glimpse of what is actually happening in the field.
- Current Sponsors: 0
- Past Sponsors: 0
- Total Sponsors: 0
- Minimum Sponsorship: $1.00
Featured Works
umitkaanusta/reddit-detective
Play detective on Reddit: Discover political disinformation campaigns, secret influencers and more
Language: Python - Stars: 214umitkaanusta/TIA
Your Advanced Twitter stalking tool
Language: Jupyter Notebook - Stars: 150umitkaanusta/Jomini
Historical battle simulation package for Python
Language: Python - Stars: 46umitkaanusta/FraudRingDetection-TrustNetworks-Trying-New-Approach
Detecting fraud rings in a p2p trust network with high AUC score (better than 9 of 10 well-known algorithms)* using basic and intuitive trust metrics
Language: Jupyter Notebook - Stars: 4Past Sponsorships
View AllSponsorship Breakdown by Kind
- User: 3