Data Science Yes, But in Kotlin!
summary
A runnable Kotlin data-science stack is presented as an alternative to Python, combining Kotlin Notebook with Maven dependencies, DuckDB accessed via JDBC, Apache Arrow for zero-copy data transfer, Kotlin Dataframe and Candy charts. The demo loads a 9 GB DuckDB dataset (~400 million taxi rows), runs a 500-row query in ~2 s and a full aggregation in ~14 s, then performs joins (including per-week joins with weather) and plots per-quarter prices and other visualizations; code is available on GitHub and interactive joins/plots work though some tools remain under development.
AI-written summary · about
chapters (6)
- 0:00Introduction to Kotlin data science
- 3:13Kotlin notebooks demo
- 11:06DataFrame data manipulation
- 19:35DuckDB and Apache Arrow
- 27:44Big data analysis and joins
- 30:13Conclusion and resources
related talks
38mData Analysis Yes, But in Kotlin!
47mData Science with Kotlin: Leveraging Deep Analysis with DataFrame
41mKotlin for Data Science
45mUsing Kotlin for Data Science
51mDataFrame: Kotlin's Innovative Approach to Data Structures
50mBuilding Data Science Workflows with Kotlin
39mKotlin Notebooks: Data Visualization for Developers (No Scientist Required)
39mReplacing SQL with Kotlin's 'dataframe' on the Las Vegas Strip
