HW 11: Data Analysis

This homework grants you a lot of freemdom. The goal is to become familiar with data analysis. Either create an ipython notebook or a plotly dashboard. For the latter you can find the tutorial here. Useful packages for data analysis include pandas, statsmodels, and for visualization matplotlib and seaborn. Note that no grader is included.

To be awarded full points, your homework needs to:

• be accompanied with all required data or the data must be available publicly on the internet,
• run with the provided input data without crashing,
• have proper documentation,
• be implemented in English including variable names, docstrings, comments, and so on.

If you need to open some files in your working folder please put these lines at the end of you script files:

if __name__ == "__main__":
    import os
    script_dir = os.path.dirname(os.path.abspath(__file__))
    os.chdir(script_dir)
    main()

Data Analysis (9 points)

Pick any data set you find interesting, and which is comprehensive enough for a meaningful analysis adressing all aspects mentioned below. A potential source for data sets is Kaggle (but you can choose any other (puclic) data source). Do a proper analysis on that data with an ipython notebook (together with pandas) or a plotly dashboard. However, please ensure that the raw data file has at most 10 megabytes.

If you create an ipython notebook, at the very least, your analysis must:

• create a nice document intermixing explanations written in Markdown and code demonstrating that you understood the basics of pandas,
• do some (basic) statistical analysis,
• perform a grouping,
• generate a plot.

If you create a plotly dashboard, at the very least, the analysis and dashboard must:

• create a nice dashboard that is easily readable without knowing the code or data using data visualization and text output (HTML) components,
• do some (basic) statistical analysis in the background and display the results
• perform a grouping,
• use at least one interactive UI component (e.g. slider, dropdown, ...),
• generate a plot.

You are free to use any data you like. Either download the data from script, or include all required files in your repository In any case, perform (and explain) your analysis such that any reader unfamiliar with the data gets a good overview about it.

NOTE: You should should present your analysis in the session after the submission deadline.

NOTE: Do not use any data set used in a tutorial, nor standard data sets like titanic and setarosa and so on.

NOTE: you can mix plotly with a notebook.

IMORTANT NOTE: make sure that you understand the data you are working with. It is useless to do some analysis without being able to interpret it.