Notebook séance 7

Chargement des données

On ajoute aussi des colonnes utiles

import pandas
import matplotlib.pyplot as plt
import matplotlib as mpl
from matplotlib import cm

df = pandas.read_excel("Donnees_M2_RD.xlsx")
df["diff_dist"] = df["Dist_A"] - df["Dist_B"]
df["reussi"] = (((df["Dist_A"] > df["Dist_B"]) & (df["Response"] == 2)) 
                 | ((df["Dist_A"] < df["Dist_B"]) & (df["Response"] == 1)))

Exercices set_index

Obtenir la DataFrame indexée par Subject et Space

df_ss = df.set_index(["Subject","Space"])
df_ss

Name_A Name_B Dist_A Dist_B Mode Side Response RT diff_dist reussi
Subject Space
P_ADI_331 E 0 2 2 4 Dic D 2 18865 -2 False
E 1 4 4 1 Dic D 2 13157 3 True
E 4 3 3 2 Dic D 1 11628 1 False
E 2 4 4 1 Dic D 1 10068 3 False
E 1 2 2 4 Dic D 1 11801 -2 True
... ... ... ... ... ... ... ... ... ... ... ...
P_VAR_330 I 0 1 3 5 Dio D 1 7626 -2 True
I 3 2 5 1 Dio D 2 6349 4 True
I 2 0 4 2 Dio D 2 9031 2 True
I 0 2 2 1 Dio D 2 16323 1 True
I 0 3 5 1 Dio D 2 10139 4 True

9594 rows × 10 columns

Obtenir la DataFrame indexée par Subject, Space et l’index original avec un niveau nommé essai (on peut utiliser .rename sur un index)

df_sse = df.set_index([
    "Subject",
    "Space",
    df.index.rename("essai")])
df_sse

Name_A Name_B Dist_A Dist_B Mode Side Response RT diff_dist reussi
Subject Space essai
P_ADI_331 E 0 0 2 2 4 Dic D 2 18865 -2 False
1 1 4 4 1 Dic D 2 13157 3 True
2 4 3 3 2 Dic D 1 11628 1 False
3 2 4 4 1 Dic D 1 10068 3 False
4 1 2 2 4 Dic D 1 11801 -2 True
... ... ... ... ... ... ... ... ... ... ... ... ...
P_VAR_330 I 9589 0 1 3 5 Dio D 1 7626 -2 True
9590 3 2 5 1 Dio D 2 6349 4 True
9591 2 0 4 2 Dio D 2 9031 2 True
9592 0 2 2 1 Dio D 2 16323 1 True
9593 0 3 5 1 Dio D 2 10139 4 True

9594 rows × 10 columns

Exercices groupby

Faire une dataframe du temps de réponse moyen par Name_A, Name_B

df.groupby(by=["Name_A", "Name_B"])["RT"].mean()

Name_A  Name_B
0       1         12232.680585
        2         12606.089770
        3         11913.616667
        4         11866.710417
1       0         11688.033333
        2         12291.152083
        3         11958.879167
        4         11601.102083
2       0         11608.336117
        1         12040.025000
        3         11978.183716
        4         11357.747917
3       0         12121.487500
        1         12884.901879
        2         12542.287500
        4         11793.595833
4       0         12300.860417
        1         11875.852083
        2         12172.350000
        3         12170.338205
Name: RT, dtype: float64

Faire une dataframe du temps de réponse max par sujet et par différence de distance (symbolique)

df.groupby(by=["Subject","diff_dist"])["RT"].max()

Subject    diff_dist
P_ADI_331  -4           16642
           -3           15652
           -2           18865
           -1           16347
            1           18067
                        ...  
P_VAR_330  -1           45627
            1           80729
            2           41271
            3           35480
            4           38072
Name: RT, Length: 192, dtype: int64

Faire une dataframe du nombre, puis du taux de bonnes réponses par Mode, Side et Space

total = df.groupby(by=["Mode", "Side", "Space"])["Subject"].count()
reussi = df[df["reussi"]].groupby(by=["Mode", "Side", "Space"])["Subject"].count()
df_mss = pandas.DataFrame(total).rename(columns={"Subject":"total"})
df_mss["reussis"] = reussi
df_mss["taux_reussis"] = df_mss["reussis"]/df_mss["total"]
df_mss

total reussis taux_reussis
Mode Side Space
Dic D E 1199 905 0.754796
I 1198 1037 0.865609
G E 1199 919 0.766472
I 1199 1040 0.867389
Dio D E 1200 824 0.686667
I 1200 940 0.783333
G E 1199 898 0.748957
I 1200 1083 0.902500

Graphiques 3D

Faire un scatter plot du taux de réussite en fonction de la différence en mètres d’une part et de l’arrondit du temps de réaction en dixièmes de seconde d’autre part.

Intégration de la différence de distance en mètres à partir du code de la séance 5

dist_i_m = pandas.DataFrame(
    { "Dist_I_m": [ 0.2, 0.3, 0.4, 0.6, 0.8 ] },
    index = [1,2,3,4,5]
)
dist_e_m = pandas.DataFrame(
    { "Dist_E_m": [ 2, 3, 4, 6, 8 ] },
    index = [1,2,3,4,5]
)
df_e = df[df['Space'] == 'E']
df1 = pandas.merge(df_e, dist_e_m, left_on='Dist_A', right_index=True)
df2 = df1.rename(columns={ 'Dist_E_m': 'Dist_A_m' })
df3 = pandas.merge(df2, dist_e_m, left_on='Dist_B', right_index=True)
df_e_m = df3.rename(columns={ 'Dist_E_m': 'Dist_B_m' })
df_i = df[df['Space'] == 'I']
df1 = pandas.merge(df_i, dist_i_m, left_on='Dist_A', right_index=True)
df2 = df1.rename(columns={ 'Dist_I_m': 'Dist_A_m' })
df3 = pandas.merge(df2, dist_i_m, left_on='Dist_B', right_index=True)
df_i_m = df3.rename(columns={ 'Dist_I_m': 'Dist_B_m' })
df_m = pandas.concat([df_e_m,df_i_m])
# Nouvelle ligne pour ajouter la différence de distances en mètres
df_m["diff_dist_m"] = df_m["Dist_A_m"] - df_m["Dist_B_m"]
df_m

Subject Name_A Name_B Dist_A Dist_B Mode Space Side Response RT diff_dist reussi Dist_A_m Dist_B_m diff_dist_m
0 P_ADI_331 0 2 2 4 Dic E D 2 18865 -2 False 3.0 6.0 -3.0
4 P_ADI_331 1 2 2 4 Dic E D 1 11801 -2 True 3.0 6.0 -3.0
7 P_ADI_331 0 3 2 4 Dic E D 1 13237 -2 True 3.0 6.0 -3.0
36 P_ADI_331 3 1 2 4 Dic E D 1 9523 -2 True 3.0 6.0 -3.0
51 P_ADI_331 2 4 2 4 Dio E G 1 11331 -2 True 3.0 6.0 -3.0
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
9498 P_VAR_330 4 2 4 3 Dic I G 2 4811 1 True 0.6 0.4 0.2
9503 P_VAR_330 3 0 4 3 Dic I G 2 6628 1 True 0.6 0.4 0.2
9526 P_VAR_330 0 1 4 3 Dic I G 2 5458 1 True 0.6 0.4 0.2
9535 P_VAR_330 2 4 4 3 Dic I G 1 8956 1 False 0.6 0.4 0.2
9546 P_VAR_330 2 3 4 3 Dio I D 2 13525 1 True 0.6 0.4 0.2

9594 rows × 15 columns

On ajoute l’arrondi au 1/10 de second de RT

df_m["RT_round"] = df_m["RT"] // 1000
df_m

Subject Name_A Name_B Dist_A Dist_B Mode Space Side Response RT diff_dist reussi Dist_A_m Dist_B_m diff_dist_m RT_round
0 P_ADI_331 0 2 2 4 Dic E D 2 18865 -2 False 3.0 6.0 -3.0 18
4 P_ADI_331 1 2 2 4 Dic E D 1 11801 -2 True 3.0 6.0 -3.0 11
7 P_ADI_331 0 3 2 4 Dic E D 1 13237 -2 True 3.0 6.0 -3.0 13
36 P_ADI_331 3 1 2 4 Dic E D 1 9523 -2 True 3.0 6.0 -3.0 9
51 P_ADI_331 2 4 2 4 Dio E G 1 11331 -2 True 3.0 6.0 -3.0 11
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
9498 P_VAR_330 4 2 4 3 Dic I G 2 4811 1 True 0.6 0.4 0.2 4
9503 P_VAR_330 3 0 4 3 Dic I G 2 6628 1 True 0.6 0.4 0.2 6
9526 P_VAR_330 0 1 4 3 Dic I G 2 5458 1 True 0.6 0.4 0.2 5
9535 P_VAR_330 2 4 4 3 Dic I G 1 8956 1 False 0.6 0.4 0.2 8
9546 P_VAR_330 2 3 4 3 Dio I D 2 13525 1 True 0.6 0.4 0.2 13

9594 rows × 16 columns

Calcul du taux de réussite

total_essais = df_m.groupby(by=["diff_dist_m", "RT_round"])["RT"].count().rename("total")
df_scatter = pandas.DataFrame(total_essais)
essais_reussis = df_m[df_m["reussi"]].groupby(by=["diff_dist_m", "RT_round"])["RT"].count().rename("reussis").reindex(index=df_scatter.index, fill_value=0)
df_scatter["reussis"] = essais_reussis
df_scatter["taux"] = essais_reussis / total_essais
fig, ax = plt.subplots(subplot_kw={'projection': '3d'})
ax.scatter(df_scatter.index.get_level_values(0), df_scatter.index.get_level_values(1), df_scatter['taux'])

<mpl_toolkits.mplot3d.art3d.Path3DCollection at 0x7fe6ca63a1f0>

png