Export figures and add additional linear regression analyses.

exploration/expl_baseline.py

@@ -89,6 +89,6 @@ baseline_inactive = baseline_inactive.assign(
 baseline_inactive["age"].describe()
 
 # %%
-3710/365.25
+3618 / 365.25
 
 # %%

statistical_analysis/adherence.py

@@ -51,25 +51,6 @@ baseline_inactive = baseline[
     baseline["Gebruikersnaam"].isin(participants_inactive_usernames)
 ]
 
-# %%
-df_esm_inactive = get_esm_data(participants_inactive_usernames)
-
-# %%
-df_esm_preprocessed = preprocess_esm(df_esm_inactive)
-df_session_counts_time = classify_sessions_by_completion_time(df_esm_preprocessed)
-
-# %%
-tbl_session_outcomes = df_session_counts_time.reset_index()[
-    "session_response"
-].value_counts()
-
-# %%
-print("All sessions:", len(df_session_counts_time))
-print("-------------------------------------")
-print(tbl_session_outcomes)
-print("-------------------------------------")
-print(tbl_session_outcomes / len(df_session_counts_time))
-
 # %%
 VARIABLES_TO_TRANSLATE = {
     "Gebruikersnaam": "username",
@@ -83,9 +64,37 @@ baseline_inactive = baseline_inactive.assign(
     age=lambda x: (now - x.date_of_birth).dt.days / 365.25245,
 )
 
+# %%
+df_esm_inactive = get_esm_data(participants_inactive_usernames)
+
+# %% [markdown]
+# # Classify EMA sessions
+
+# %%
+df_esm_preprocessed = preprocess_esm(df_esm_inactive)
+df_session_counts_time = classify_sessions_by_completion_time(df_esm_preprocessed)
+
+# %% [markdown]
+# Sessions are now classified according to the type of a session (a true questionnaire or simple single questions) and users response.
+
 # %%
 df_session_counts_time
 
+# %%
+tbl_session_outcomes = df_session_counts_time.reset_index()[
+    "session_response"
+].value_counts()
+
+# %%
+print("All sessions:", len(df_session_counts_time))
+print("-------------------------------------")
+print(tbl_session_outcomes)
+print("-------------------------------------")
+print(tbl_session_outcomes / len(df_session_counts_time))
+
+# %% [markdown]
+# ## Consider only true EMA sessions
+
 # %%
 df_session_finished = df_session_counts_time[
     df_session_counts_time["session_response"] == SESSION_STATUS_COMPLETE
@@ -117,6 +126,9 @@ df_adherence
 # %%
 df_adherence.describe()
 
+# %%
+df_adherence[["gender", "startlanguage"]].value_counts()
+
 # %%
 sns.displot(df_adherence["finished_sessions"], binwidth=5, height=5)
 
@@ -209,7 +221,9 @@ df_session_workday = df_session_workday.assign(
 )
 
 # %%
-sns.displot(df_session_workday["time_diff_minutes"], binwidth=5, height=5)
+g1 = sns.displot(df_session_workday["time_diff_minutes"], binwidth=5, height=5)
+g1.set_axis_labels("Time difference [min]", "Session count")
+# g1.savefig("WorkdayEMAtimeDiff.pdf")
 
 # %% [markdown]
 # There are some sessions that are really close together. By design, none should be closer than 30 min. Let's take a look at those.
@@ -246,6 +260,11 @@ df_session_workday = df_session_workday[df_session_workday.time_diff_minutes > 2
 # %%
 df_session_workday.describe()
 
+# %%
+df_session_workday[df_session_workday["time_diff_minutes"] < 120].shape[
+    0
+] / df_session_workday.shape[0]
+
 # %% [markdown]
 # These statistics look reasonable.
 
@@ -253,13 +272,27 @@ df_session_workday.describe()
 # ### Differences between participants
 
 # %%
-df_mean_daytime_interval = df_session_workday.groupby("participant_id").mean()
+df_mean_daytime_interval = df_session_workday.groupby("participant_id").median()
 
 # %%
 df_mean_daytime_interval.describe()
 
 # %%
-sns.displot(df_mean_daytime_interval.time_diff_minutes, binwidth=5, height=5)
+g2 = sns.displot(df_mean_daytime_interval.time_diff_minutes, binwidth=5, height=5)
+g2.set_axis_labels("Median time difference [min]", "Participant count")
+# g2.savefig("WorkdayEMAtimeDiffMedianParticip.pdf")
+
+# %%
+df_adherence = df_adherence.merge(
+    df_mean_daytime_interval, how="left", left_on="participant_id", right_index=True
+)
+
+# %%
+lr_ols_time_diff_median = smf.ols(
+    "time_diff_minutes ~ C(gender) + C(startlanguage) + age", data=df_adherence
+)
+ls_result_time_diff_median = lr_ols_time_diff_median.fit()
+ls_result_time_diff_median.summary()
 
 # %%
 df_count_daytime_per_participant = df_session_workday.groupby(
@@ -288,6 +321,9 @@ s_evening_completed = df_session_counts_time_completed.groupby(
 # %%
 df_session_counts_time_completed
 
+# %%
+s_evening_completed.sum()
+
 # %%
 s_evening_completed_ratio = (
     s_evening_completed.groupby("participant_id").sum()
@@ -298,6 +334,23 @@ s_evening_completed_ratio = (
 s_evening_completed_ratio.describe()
 
 # %%
-sns.displot(s_evening_completed_ratio - 0.001, binwidth=0.1, height=5)
+g3 = sns.displot(s_evening_completed_ratio - 0.001, binwidth=0.05, height=5)
+g3.set_axis_labels("Ratio of days with the evening EMA filled out", "Participant count")
+# g3.savefig("EveningEMAratioParticip.pdf")
+
+# %%
+df_adherence = df_adherence.merge(
+    s_evening_completed_ratio.rename("evening_EMA_ratio"),
+    how="left",
+    left_on="participant_id",
+    right_index=True,
+)
+
+# %%
+lr_ols_evening_ratio = smf.ols(
+    "evening_EMA_ratio ~ C(gender) + C(startlanguage) + age", data=df_adherence
+)
+ls_result_evening_ratio = lr_ols_evening_ratio.fit()
+ls_result_evening_ratio.summary()
 
 # %%