Exploratory data analysis on National Highway Transportation Safety Association (NHTSA) crash test results
The goal of this project was to gain experience in data cleaning and preliminary EDA, using data science development tools. EDA concluded that missing data and predominance of ordinal data (star ratings) make additional testing difficult. This may not be the ideal dataset for additional testing and modeling.
Next, we'll look for a cleaner dataset with more opportunities for end-to-end analysis.
- Configure development environment: VS Code, Python, Jupyter Notebooks.
- Find an interesting dataset for analysis.
- Begin exploratory data analysis: distributions, missing values, descriptive stats, etc.
- Consider inferential statistics. Look for patterns to build and test hypotheses.
- Dataset: NHTSA Safercar Dataset
- Visual Studio Code
- Jupyter Notebook
- Python: Pandas, Pyplot, Seaborn