Focus

Use simple libraries and functions to do explorary data analysis.

Programming language

Python

Dataset 1: 'ri_police.csv'

• Traffic stops by police officers
• Collected data from 31 US states, but only 'Rhode Island' in this case
• Some of the columns and rows have been removed, full dataset can be downloaded at https://openpolicing.stanford.edu/

Dataset 2: 'Weather in ri.csv'

• Exploring the weather dataset
• From NOAA: National Centers For Environmental Information

Libraries

• pandas
• matplotlib.pyplot

Functions

• df.isnull().sum()
• df.shape
• df.dropna(subset=[' '], inplace=True)
• df.col.astype(' ') / df.dtypes
• df.col1.str.cat(ri.col2, sep=' ')
• pd.to_datetime(col)
• df.value_counts() / df.value_counts(normalize=True)
• df.col.mean()
• df.groupby(['col1', 'col2']).col3.mean()
• df['bool column'] = df.col1.str.contains('string', na=False)
• df.plot() / df.plot(subplots=True) / (kind='bar', stacked= True) / (kind='barh') / (kind='hist', bins=n) / (kind='box')
• plt.xlabel(' ') / plt.ylabel(' ') / plt.title(' ') / plt.show()
• df.resample(' ').mean()
• pd.crosstab(df.col1, df.col2)
• df.loc['col1': 'col2'])
• df.col.unique()
• df.col.map({org1 : map1, org2 : map2,.....})
• df.sort_values()
• df['col1', 'col2',...].describe()
• df.sum(axis='columns')
• df.reset_index(inplace=True)
• df.set_index('col', inplace=True)
• newdf = pd.merge(left=df1, right=df2, left_on='df1.col1', right_on='df2.col2', how='left') #right
• df.pivot_table(index='col1', columns='col2', values='col3')