From 2103716f791d8c082237a38f8a03c5b74781c671 Mon Sep 17 00:00:00 2001 From: Daniel Svitan Date: Sat, 21 Dec 2024 19:41:53 +0100 Subject: [PATCH] :zap: Adds commute and living violin plot --- analyze_commute.py | 19 ++++++++++++------- analyze_living.py | 20 +++++++++++++------- 2 files changed, 25 insertions(+), 14 deletions(-) diff --git a/analyze_commute.py b/analyze_commute.py index f0bd297..6bb1328 100644 --- a/analyze_commute.py +++ b/analyze_commute.py @@ -1,6 +1,6 @@ import numpy as np -from analyze import analyze +from analyze import analyze, plot_violin dataset = np.load("clean.npy") print(f"dataset shape: {dataset.shape}; analyzing column 9 (commute)") @@ -14,16 +14,21 @@ print("") def analyze_living(name: str, col: np.ndarray): occupation_col = dataset[:, 9] - analyze(name, [ + data = [ col[occupation_col == 0], col[occupation_col == 1], col[occupation_col == 2], col[occupation_col == 3], col[occupation_col == 4] - ]) + ] + F, p = analyze(name, data) + return data, F, p -analyze_living("gpa", dataset[:, 2]) -analyze_living("math", dataset[:, 3]) -analyze_living("slovak", dataset[:, 4]) -analyze_living("english", dataset[:, 5]) +data_gpa, F_gpa, p_gpa = analyze_living("gpa", dataset[:, 2]) +data_math, F_math, p_math = analyze_living("math", dataset[:, 3]) +data_slovak, F_slovak, p_slovak = analyze_living("slovak", dataset[:, 4]) +data_english, F_english, p_english = analyze_living("english", dataset[:, 5]) + +plot_violin([data_gpa, data_math, data_slovak, data_english], ["Intrák", "<= 15m", "<= 30m", "<= 1h", "> 1h"], + [F_gpa, F_math, F_slovak, F_english], [p_gpa, p_math, p_slovak, p_english], "Dochádzanie") diff --git a/analyze_living.py b/analyze_living.py index 567dcf1..bdac044 100644 --- a/analyze_living.py +++ b/analyze_living.py @@ -1,6 +1,6 @@ import numpy as np -from analyze import analyze +from analyze import analyze, plot_violin dataset = np.load("clean.npy") print(f"dataset shape: {dataset.shape}; analyzing column 8 (living)") @@ -14,16 +14,22 @@ print("") def analyze_living(name: str, col: np.ndarray): occupation_col = dataset[:, 8] - analyze(name, [ + data = [ col[occupation_col == 0], col[occupation_col == 1], col[occupation_col == 2], col[occupation_col == 3], col[occupation_col == 4] - ]) + ] + F, p = analyze(name, data) + return data, F, p -analyze_living("gpa", dataset[:, 2]) -analyze_living("math", dataset[:, 3]) -analyze_living("slovak", dataset[:, 4]) -analyze_living("english", dataset[:, 5]) +data_gpa, F_gpa, p_gpa = analyze_living("gpa", dataset[:, 2]) +data_math, F_math, p_math = analyze_living("math", dataset[:, 3]) +data_slovak, F_slovak, p_slovak = analyze_living("slovak", dataset[:, 4]) +data_english, F_english, p_english = analyze_living("english", dataset[:, 5]) + +plot_violin([data_gpa, data_math, data_slovak, data_english], + ["Rodina", "Príslušník/čka", "Sám/a | spolu", "Intrák", "Iné"], [F_gpa, F_math, F_slovak, F_english], + [p_gpa, p_math, p_slovak, p_english], "Životná situácia")