Foundations of Data Science_Week 7 Notes

Foundations of Data Science Mar 19, 2023 Module 6: Descriptive Statistics and Visualization - https://colab.research.google.com/drive/1jbT-MH_9tSA1NG_TED_yXvvAePZZimzO?usp=sharing - To important mat plot library, use the following code: import.matplot.lib.pyplot as plt Plotting with Pandas - To create a histogram , use the following code: ‘tablename.hist(bins=50, figsize=(20,15)); o Will create a histogram for every numerical column of data o Will create single data histograms - To customize a histogram , like change color, specific column data or block size, you can use the following code: tablename[“column name’].plot(kind=’hist’, figsize=(15,10), title=”column name”, color=”color”) o Example: df_housing["median_house_value"].plot(kind=’hist’, figsize=(15,10), title="Median House Value", color="green") - To save an image, use the following code: plt.gcf.savefig(“name of the image.png”) - To make a stacked histogram , use the following steps: o 1. Make a pivot table of the data columns you want to add to the stacked histogram table.pivot(columns=”column 1”)[“column 2”] o 2. Use the histogram code to create the graph table.pibot(columns=”column 1”)[“column 2”].plot(kind=’hist’, stacked = True, figsize =(15,10), title=”title of your graph”); - To make a bar plot , follow the steps below: o 1. Make a new table of the information you want to graph o 2. Use the following code to make a bar plot: ax = newtable.plot(kind=’bar’, rot = 45, figsize=(15,10)); ax.set_ylabel(“name of y axis”); ax.set+xlabel(“name of x axis”);  Rot parameter rotates the labels - To make a stacked bar plot , use the following steps o 1. Make a new crosstab table with the information you want to graph df_crosstab = pd.crosstab(tablename["column 1"], table name["column 2"]) o 2. Make a stacked bar graph: ax = df.crosstab.plot(kind=’bar’, stacked = True, rot = 0, figsize=(15,10)); ax.set_ylabel(“name of y axis”); - To make a grouped bar plot , follow the steps below: o 1. Make a new table with the information you want to display df.newtable = table.groupby(“column 1”)[“column 2”, “column 3”]].function().reset_index()  An example of function could be median() or mean()

Foundations of Data Science Mar 19, 2023 o 2. Make a grouped bar graph ax.df.newtable.plot(kind=’bar’, x=”column 1”, y=[“column 2”, “column 3”], rot = 0, figsize=(15,10)); ax.set_ylabel(“name of y axis”)  Column 2 and 3 will be the two columns you want the grouped bar graph to show on the x axis - To make a scatter plot , use the following code: o Table.plot(kind=”scatter”, x=“column”, y=”column”, figsize=(15,10), s=2, alpha=0.8, c=”color of your choice”);  S -> size of the dots on the graph  Alpha -> how transparent the dots are on the graph o To zoom in on a scatter plot, add specific x-axis and y-axis ranges ax.set_xlim((#,#)) -> it will only show the data between these values on the x and y axis ax.et_ylim((#,#)) - To graph a bubble plot , look at the workbook (link at the top of document) - To make a line plot , use the following steps: o 1. Make a new table with all the information you want to graph o 2. Use the code to make a line plot: ax.df.newtable.plot(figsize=(15,10), linewidth=4) ax.set_ylabel(“name of y axis): ax.set_xlabel(“name of x axis”); plt.legend(loc=”upper/lower right/left”); -> use loc function to determine where your legend will appear on the table - To create a two line plot on the same axis, use the following code: o 1. Ax = (table[“column”].sort_values().reset_index(drop=True).plot(figsize=(15,10),linewidth=4,l abel=”name of line”); table[“column”].sort_values().reset_index(drop=True).plot(ax=ax, linewidth=4, label=”name of line”); ax.set_ylabel(“name of y axis”); ax.set_xaxis(“name of x axis”); plt.legend(loc=”location”); o Have to ensure that both lines are using the same x- axis as a reference line Plotting with Matplotlib - Histogram : plt.figure(figsize=(10,8)) plt.hist(table[“column”]) plt.title(“title of graph”) plt.ylabel(“name of y axis”) plt.xlabel(“name of x axis”) plt.savefig(“name of file.png”, dpi=180)