Mirror Mirror Study - 1

Before the study session, everyone agreed to study Chapter 1 in advance, and 2 hours every Tuesday were set aside for those who had trouble keeping up. Participants created GitHub accounts, learned how to use git, and also installed IPython Notebook to prepare a hands-on practice environment.

Rather than starting with a difficult approach, the group began with the mindset of treating the tools lightly, like toys. That's how the study started, and the direction was set for everyone to proceed together, keeping pace with one another.

During the summary process, methods for selecting specific columns were organized.

• Using column index values: makes it easy to know column positions, and is suitable when column positions don't change even while processing multiple input files.
• Using column headers: makes column names easy to identify, and is useful when column positions differ across files but the headers are the same.

In the advanced section, the advantages of IPython Notebook were reconfirmed. Since results can be seen immediately, experimentation and iteration are easy, and you can learn quickly by running code with Shift+Enter or checking function arguments with Shift+Tab. It was also emphasized to try things yourself when unsure, then check the help, and search in English to refer to official documentation.

pandas was introduced as a tool that borrows syntax from R to make data analysis easier. For data selection, it was summarized that you can use .iloc for position-based selection and .loc for label-based selection, and while .ix is used for mixed indexing, it's better to explicitly use .loc or .iloc unless there's a special case.

Finally, it was concluded that quickly experimenting with IPython Notebook while learning basic Python syntax together with pandas is effective. The key point was that even with data files containing many columns, analysis efficiency can be improved by handling only the necessary columns.