INTRO: For this project, we start from the provided database, which is available from Global Shark Attacks ๐.
OBJECTIVE: The purpose of the exercise is to be able to put into practice everything we have learned about data exploration, cleaning, analysis and visualization.
REQUIREMENTS: We must establish at least one hypothesis on the basis of which we will clean the dataset. Make at least two graphs that support those assumptions. Use at least 5 different data cleaning techniques. HYPOTHESIS:
- The great white shark is the most lethal shark.
- He likes to eat USA people more than anything, so he goes an extra mile to kill them.
- He likes to eat surfers more than to eat swimmers.
ORGANIZATION:
- Cleanup:
After a brief exploration of the size and quality of the data, I proceed to a first basic roughing out. In the following phases I went through the file by columns, evaluating the best way to extract the most data according to the content, the type of data and the usefulness it could provide us for future visualizations. This process is documented in: clean.ipynb commented file with the cleaning process.
cleaning_functions.py functions created ad hoc for cleaning and extraction.
clean_sharks.csv file obtained in the last step after the cleaning process.
- Analysis:
Once the clean file has been obtained, several graphs have been made, relying on different libraries, in order to visually support the conclusions. This process is documented in: EDA.ipynb file that includes both the graphs and the conclusions obtained from the study.