Please see the inital version of this analysis in 2 parts (Jupyter notebooks using various Python 3 libraries).
-
part1 (Part 1 Notebook - download data for analysis via sports API - NBA V2)
-
Raw data pulled from NBA sports API calls available at
/raw
dir with all teams/seasons data, please contactconorheffron
to collaborate on this analysis etc. -
part2-V1 (Part 2 Notebook - read, clean, analyse, & visualise raw data stored on disk)
-
part1-V2 (Part 1 Notebook - download data for analysis via sports API - NBA V2)
-
Raw data pulled from NBA sports API calls available at
/raw
dir with all teams/seasons data, please contactconorheffron
to collaborate on this analysis etc. -
part2-V2 (Part 2 Notebook - read, clean, analyse, & visualise the raw data stored on disk)
Renewed versions including 2024 NBA games/players data & the introduction of prompt engineering with Open AI Large Language Models (LLMs) using BambooLLM & GPT 3.5 Turbo.
-
part1-V3 (Part 1 Notebook - download data for analysis via sports API - NBA V2)
-
The data pulled from NBA sports API calls is dumped & stored at
./raw
dir with all teams/seasons data, please contactconorheffron
to collaborate on this analysis etc. -
part2-download-data-V3 (Part 2 Notebook - Read & clean raw data stored on disk then download processed data into single CSV file)
-
Intermediary file format to speed up development/analysis phase available in root folder at
./nba-stats-data.csv
. This file is produced by running part 2 notebook. -
part3-analysis-V3 (Part 3 Notebook - Analyse & visualise pre-processed data stored in single CSV file from part 2 notebook)
-
part4-ai-llm-analysis-V3 (Part 4 Notebook - Speak to the data via Large Language Model (BambooLLM) prompt that uses the NBA Stats CSV file generated by part 2 notebook)
-
part4-ai-llm-analysis-V3-GPT3-5-Turbo (Part 4 Notebook - Speak to the data via Large Language Model (GPT 3.5 Turbo) prompt that uses the NBA Stats CSV file generated by part 2 notebook)
- part4-ai-llm-analysis-V4-GPT-4 (Part 4 Notebook - Speak to the data via Large Language Model (GPT 4) prompt that uses the NBA Stats CSV file generated by part 2 notebook)