heatmap

Similarity, matrix, and confusion plots.

df = sns.load_dataset('titanic').dropna(subset=['age', 'fare', 'class', 'sex', 'survived']).reset_index(drop=True)
df2 = df[['age', 'fare', 'sibsp', 'parch']].head(8).copy()
df2.index = [f'row_{i}' for i in range(len(df2))]
print(df.shape)
print(df2.shape)

(714, 15)
(8, 4)

df.head()

	survived	pclass	sex	age	sibsp	fare	embarked	class	who	adult_male	deck	embark_town	alive	alone
0	0	3	male	22.0	1	7.2500	S	Third	man	True	NaN	Southampton	no	False
1	1	1	female	38.0	1	71.2833	C	First	woman	False	C	Cherbourg	yes	False
2	1	3	female	26.0	0	7.9250	S	Third	woman	False	NaN	Southampton	yes	True
3	1	1	female	35.0	1	53.1000	S	First	woman	False	C	Southampton	yes	False
4	0	3	male	35.0	0	8.0500	S	Third	man	True	NaN	Southampton	no	True

Matrix Helpers

get_similarity


def get_similarity(
    df:DataFrame, # numeric feature matrix indexed by sample name
    metric:str='euclidean', # pairwise_distances metric name
)->tuple:

Calculate both distance and similarity matrices for a dataframe.

get_similarity(df2)[0]

	row_0	row_1	row_2	row_3	row_4	row_5	row_6	row_7
row_0	0.000000	66.001996	4.177993	47.657345	13.062925	54.911521	24.415786	6.714166
row_1	66.001996	0.000000	64.492435	18.429118	63.312323	25.182682	61.821302	61.188418
row_2	4.177993	64.492435	0.000000	46.073643	9.000868	52.100901	27.548548	3.910651
row_3	47.657345	18.429118	46.073643	0.000000	45.061097	19.066500	46.039121	42.780883
row_4	13.062925	63.312323	9.000868	45.061097	0.000000	47.754949	35.618122	8.803791
row_5	54.911521	25.182682	52.100901	19.066500	47.754949	0.000000	60.513388	48.906725
row_6	24.415786	61.821302	27.548548	46.039121	35.618122	60.513388	0.000000	27.089433
row_7	6.714166	61.188418	3.910651	42.780883	8.803791	48.906725	27.089433	0.000000

plot_corr


def plot_corr(
    df_corr:DataFrame, # correlation, distance, or similarity matrix
    mask_method:str | None='upper', # upper, lower, or None
    inverse_color:bool=False, # reverse the colormap when True
    figsize:tuple=(15, 10), # figure size in inches
    annot:bool=True, # whether to annotate the matrix values
    linewidths:float=0.1, # cell border width
    kwargs:VAR_KEYWORD
):

Plot a square matrix with an optional triangular mask.

plot_corr(df[['age', 'fare', 'sibsp', 'parch']].corr(numeric_only=True))

Classification And Composition

plot_confusion_matrix


def plot_confusion_matrix(
    target, # true labels
    pred, # predicted labels
    class_names:list[str] | None=None, # labels shown on the axes
    normalize:bool=False, # normalize rows when True
    title:str='Confusion matrix', # plot title
    cmap:LinearSegmentedColormap=<matplotlib.colors.LinearSegmentedColormap object at 0x7f9d7e4ea5d0>, # matplotlib colormap
    figsize:tuple=(6, 6), # figure size in inches
    kwargs:VAR_KEYWORD
):

Plot a confusion matrix from target and prediction arrays.

plot_confusion_matrix(df['survived'], df['adult_male'], class_names=['False', 'True'], normalize=True)

Normalized confusion matrix