globem-user-analysis

name

globem-user-analysis

description

Comprehensive individual-user analysis on the GLOBEM dataset — a longitudinal passive-sensing + mental-health study of college students. Use this skill whenever a task involves analyzing a specific participant (e.g. "Analyze user INS-W_002") from the GLOBEM dataset, exploring behavioral patterns from smartphone sensors, correlating behavioral signals with mental health outcomes, or producing a comprehensive user profile from multimodal sensing data.

GLOBEM Individual User Analysis

Dataset Overview

The GLOBEM dataset tracks college students over a 92-day period (~April–July) using passive smartphone sensing. Each participant has:

Sensor CSVs (daily rows, columns: Unnamed: 0, pid, date, + features):

File	Modality	Key Columns
activity_allday_raw.csv	Steps, sedentary/active bouts	intraday_rapids_sumsteps, intraday_rapids_countepisodesedentarybout, intraday_rapids_countepisodeactivebout
sleep_allday_raw.csv	Duration, efficiency, timing	summary_rapids_avgdurationasleepmain, summary_rapids_avgefficiencymain, summary_rapids_avgdurationtofallasleepmain
communication_allday_raw.csv	Call counts, duration, contacts	rapids_outgoing_count, rapids_incoming_count, rapids_missed_count, rapids_outgoing_meanduration, rapids_outgoing_distinctcontacts
connectivity_allday_raw.csv	Bluetooth scans, unique devices	rapids_countscans, rapids_uniquedevices
location_allday_raw.csv	Mobility, home time, entropy	barnett_avgflightdur, barnett_avgflightlen, barnett_homelabel, barnett_circdnrtn, barnett_rog, barnett_siglocsvisited, barnett_disttravelled
phone_usage_allday_raw.csv	Unlock frequency, usage duration	rapids_countepisodeunlock, rapids_sumdurationunlock, rapids_sumdurationunlockhome

Mental-health / survey files (read directly via execute_code; get_field_description will fail on these):

File	Contents
dep_weekly.csv	Columns: pid, date, feel_anxious, feel_depressed, BDI2, dep, dep_weekly_subscale, anx_weekly_subscale
dep_endterm.csv	End-of-study BDI2 score and dep flag
ema.csv	Daily negative_affect_EMA scores
pre.csv	Pre-study surveys (see exact columns below)
post.csv	Post-study surveys (see exact columns below)
platform.csv	iOS vs Android

Exact Pre/Post Survey Column Names

pre.csv (suffix _PRE): UCLA_10items_PRE, SocialFit_PRE, 2waySSS_receiving_emotional_PRE, 2waySSS_giving_emotional_PRE, 2waySSS_giving_instrumental_PRE, 2waySSS_receiving_instrumental_PRE, ERQ_reappraisal_PRE, ERQ_suppression_PRE, BRS_PRE, CHIPS_PRE, PSS_10items_PRE, STAIS_PRE, MAAS_7items_PRE, CESD_9items_PRE, CESD_10items_PRE, BFI10_extroversion_PRE, BFI10_agreeableness_PRE, BFI10_conscientiousness_PRE, BFI10_neuroticism_PRE, BFI10_openness_PRE

post.csv (suffix _POST; BFI10 not in POST): UCLA_10items_POST, SocialFit_POST, 2waySSS_receiving_emotional_POST, 2waySSS_giving_emotional_POST, 2waySSS_giving_instrumental_POST, 2waySSS_receiving_instrumental_POST, ERQ_reappraisal_POST, ERQ_suppression_POST, BRS_POST, CHIPS_POST, PSS_10items_POST, STAIS_POST, MAAS_7items_POST, CESD_9items_POST, CESD_10items_POST

Scale meanings (higher = more unless noted):

• STAIS — state anxiety (higher = more anxious)
• PSS_10items — perceived stress (higher = more stressed)
• CESD_9items / CESD_10items — depression symptoms (higher = worse)
• UCLA_10items — loneliness (higher = more lonely)
• BRS — resilience (higher = more resilient, good)
• ERQ_reappraisal — cognitive reappraisal use (higher = more adaptive coping)
• ERQ_suppression — emotional suppression (higher = more suppression)
• CHIPS — health stressors count (higher = more stressed)
• MAAS_7items — mindfulness (higher = more mindful, good)
• SocialFit — social fit (higher = better fit)
• 2waySSS_receiving/giving_emotional — emotional social support exchange
• 2waySSS_receiving/giving_instrumental — instrumental social support exchange
• BFI10_* — Big Five personality (extroversion, agreeableness, conscientiousness, neuroticism, openness)

Analysis Pipeline

Phase 1 — Orientation (1–2 calls)

# 1. list_files to confirm available files
# 2. get_field_description on 2-3 sensor files to learn any extra column names needed
# Do NOT call get_field_description for dep_weekly, ema, pre, post, dep_endterm, platform

Phase 2 — Per-modality stats (one call per modality)

Filter all sensor DFs by pid == '<user_id>'. For each modality compute:

• Mean ± std, min/max over valid (non-NaN) rows
• Count of valid days (report as n/92)
• Weekday vs. weekend difference (add df['date'].dt.dayofweek → 0–4 weekday, 5–6 weekend)
• Temporal trend: split by study midpoint → first-half vs. second-half means

user_df['date'] = pd.to_datetime(user_df['date'])
mid = user_df['date'].min() + (user_df['date'].max() - user_df['date'].min()) / 2
first_half = user_df[user_df['date'] < mid]
second_half = user_df[user_df['date'] >= mid]

Activity — intraday_rapids_sumsteps, intraday_rapids_countepisodesedentarybout, intraday_rapids_countepisodeactivebout

Sleep — summary_rapids_avgdurationasleepmain (minutes), summary_rapids_avgefficiencymain (already 0–100, NOT decimal — never multiply by 100), summary_rapids_avgdurationtofallasleepmain. Convert bedtime/wake minutes-since-midnight to HH:MM for readability.

Communication — rapids_outgoing_count, rapids_incoming_count, rapids_missed_count, rapids_outgoing_meanduration, rapids_outgoing_distinctcontacts. Compute outgoing/incoming ratio (>1 = proactive).

Location — barnett_disttravelled, barnett_rog, barnett_homelabel (minutes/day at home), barnett_circdnrtn (0–1 circadian consistency), barnett_siglocsvisited. Filter GPS outliers: drop values > median × 10 for distance and rog before averaging.

Phone Usage — rapids_countepisodeunlock, rapids_sumdurationunlock (minutes), rapids_sumdurationunlockhome. Report home-use fraction.

Connectivity — rapids_countscans, rapids_uniquedevices.

Phase 3 — Mental health profile (1–2 calls)

dep_weekly = pd.read_csv('dep_weekly.csv')
dep_endterm = pd.read_csv('dep_endterm.csv')
ema = pd.read_csv('ema.csv')
pre = pd.read_csv('pre.csv')
post = pd.read_csv('post.csv')
platform = pd.read_csv('platform.csv')

uid = '<user_id>'
user_dep = dep_weekly[dep_weekly['pid'] == uid].copy()
user_dep['date'] = pd.to_datetime(user_dep['date'])
user_endterm = dep_endterm[dep_endterm['pid'] == uid]
user_ema = ema[ema['pid'] == uid].copy()
user_ema['date'] = pd.to_datetime(user_ema['date'])
user_pre = pre[pre['pid'] == uid]
user_post = post[post['pid'] == uid]
user_platform = platform[platform['pid'] == uid]

Extract and report:

• Platform (iOS/Android)
• Depression: weekly flag rate, end-term BDI2 + dep status
• EMA: mean, std, min/max, first-half vs. second-half trend
• Pre→Post survey changes for ALL key scales using exact column names above (report Pre value, Post value, and change with ↑↓ arrows). Interpret direction: improved or worsened.
• Personality (BFI10 pre only)

Phase 4 — Cross-modal correlation & synthesis (1–2 calls)

This phase drives the most analytically rich insights. Complete all sub-analyses:

1. EMA ↔ Behavioral correlations (Pearson r): Merge EMA with each sensor modality on pid + date, then correlate negative_affect_EMA with:

• intraday_rapids_sumsteps (activity)
• summary_rapids_avgdurationasleepmain (sleep duration)
• barnett_homelabel (home time)
• rapids_sumdurationunlock (screen time)
• intraday_rapids_countepisodesedentarybout (sedentary behavior)

2. High vs. low EMA day comparisons: Split days by EMA median → report behavioral metric means for each group:

ema_median = user_ema['negative_affect_EMA'].median()
high_ema_dates = user_ema[user_ema['negative_affect_EMA'] > ema_median]['date']
low_ema_dates = user_ema[user_ema['negative_affect_EMA'] <= ema_median]['date']
# Compare steps, screen time, home time, etc. on high vs. low EMA days

3. Depression-flagged week behavior comparison: Aggregate daily sensor data to weekly means (group by week). Join on dep_weekly dates (±7 days), then compare depressed vs. non-depressed week means for steps, sleep duration, phone unlocks, home time.

4. Temporal trends across modalities: Summarize first-half → second-half changes for all key metrics in a consolidated table (steps, sleep, phone unlocks, distance, EMA).

Phase 5 — Data quality check (integrate into output)

For each modality, report valid days as n/92. Flag modalities with <20% coverage as "critically sparse — interpret with caution."

Synthesis Template

## Comprehensive Analysis of User <pid>

### Study Context
- Platform, study period (date range), data completeness per modality (n/92 days)

### Physical Activity
- Steps (mean ± std, min, max, valid days), sedentary/active balance, weekday vs. weekend
- Temporal trend: first-half → second-half step change

### Sleep
- Duration (hours), efficiency (%), timing (bedtime HH:MM, wake HH:MM), variability
- Temporal trend: sleep duration first-half → second-half

### Communication
- Call frequency, outgoing/incoming ratio, social diversity (distinct contacts)
- Temporal trend: outgoing call count first-half → second-half

### Location & Mobility
- Daily distance, home time (hours/day, %), circadian routine score, radius of gyration
- Temporal trend: distance or home time first-half → second-half

### Phone Usage
- Unlock count, screen time (hours), home vs. elsewhere split
- Temporal trend: unlock count or duration first-half → second-half

### Social Proximity (Connectivity)
- BT scan rate, unique devices per day

### Mental Health
- Depression trajectory: weekly flag rate (n/total weeks), end-term BDI2 + dep status
- EMA negative affect: mean ± std, trend (first-half → second-half mean)
- Pre→Post survey changes (all key scales with ↑↓ changes and interpretation)

### Cross-Modal Patterns
- EMA correlations with behavioral signals (list r values)
- Behavioral differences on high vs. low negative affect days
- Behavioral differences in depressed vs. non-depressed weeks
- Notable first-half → second-half behavioral shifts

### User Profile
- 3–5 sentence synthesis connecting behavioral patterns, temporal trends, and mental health

Common Pitfalls

1. Sleep efficiency: summary_rapids_avgefficiencymain is already percentage (e.g., 93.5). Never multiply by 100.

2. Survey columns must use exact names with _PRE/_POST suffix. Column discovery via df.columns.tolist() recommended on first access; get_field_description does NOT work on survey files.

3. Minute encoding: Bedtime/wake times are often minutes-since-midnight. Convert: f"{int(m//60):02d}:{int(m%60):02d}". Values like 1500 → 25:00 means 1:00 AM next day.

4. Sparse data: Some users have very few valid rows (<14/92 days for some modalities). Always check df[col].notna().sum() before computing stats and note effective n.

5. Location GPS outliers: barnett_disttravelled and barnett_rog can have extreme values from GPS errors. Use values[values < values.median() * 10] filtering before averaging.

6. Weekly vs. daily merge: dep_weekly is weekly; sensor data is daily. To compare, aggregate daily data into 7-day windows aligned with each dep_weekly date row.

7. EMA correlation requires merging on date: User's EMA may not overlap with all sensor dates. Use inner join and report n for each correlation.

8. 2waySSS and ERQ are often missed but frequently asked about: Always extract emotion regulation (ERQ_reappraisal/suppression) and social support exchange (2waySSS giving/receiving) from pre/post surveys.