******* 2020/2019-2018 데이터 업데이트: 2023. 2. 2 ******************************************
1. 실험자의 인구통계학적 정보가 수정되었습니다. (성별, 나이, 주사용 손, 키, 몸무게)
******* 2020/2019-2018 데이터 업데이트: 2022. 3. 17 ******************************************
1. 실험자의 인구통계학적 정보가 추가되었습니다. (성별, 나이, 주사용 손, 키, 몸무게)
2. 매일 오전 실험 시작 시 (AM), 그리고 실험 종료 시 (PM) 입력한 수면 관련 설문조사의 결과가 추가되었습니다.
3. Withings Sleep Mat (2020), Actigraph (2019), Fitbit Versa (2018)로 수집한 수면 측정 데이터가 추가되었습니다.
아래 본문의 설명을 참고하시기 바랍니다.
*******************************************************************************************************
일상생활 중 다양한 경험상황을 이해하기 위해 멀티모달 센서를 활용한 라이프로그 데이터셋을 구축하였습니다.
2020년 22명, 2019년 20명, 2018년 30명의 실험자로부터 총 1,285일의 실험일 동안 약 14,220시간의 데이터를 수집하였습니다.
데이터셋은 스마트폰의 IMU 및 GPS 데이터, E4의 가속도계, PPG, EDA, 서모파일 센서로부터 측정한 다양한 생리반응 신호,
그리고 사용자가 직접 입력한 행동, 환경, 및 감정 레이블을 포함합니다.
* 데이터셋의 구성
==================================================================================================
+----- USER_ID
| +----- Unix_epoch_timestamp (DAY 1)
| | +----- e4Acc
| | | timestamp (e4_accelerometer_data).csv
| | | ...
| | +----- e4Bvp
| | | timestamp (e4_blood_volume_pressure_data).csv
| | | ...
| | +----- e4Eda
| | | timestamp (e4_electrodermal_activity_data).csv
| | | ...
| | +----- e4Hr
| | | timestamp (e4_heart_rate_data).csv
| | | ...
| | +----- e4Temp
| | | timestamp (e4_skin_temperature_data).csv
| | | ...
| | +----- mAcc
| | | timestamp (mobile_accelerometer_data).csv
| | | ...
| | +----- mGps
| | | timestamp (mobile_gps_data).csv
| | | ...
| | +----- mGyr
| | | timestamp (mobile_gyroscope_data).csv
| | | ...
| | +----- mMag
| | | timestamp (mobile_magnetometer_data).csv
| | | ...
| | timestamp_label.csv
| +----- Unix_epoch_timestamp (DAY 2)
| | +----- ...
==================================================================================================
* ETRI_Lifelog_Dataset_2020
2020 데이터셋은 다음의 데이터를 포함합니다.
==================================================================================================
- Triaxial acceleration force (in m/s^2) from the mobile phone accelerometer (30 Hz) and E4 accelerometer (32 Hz)
- Triaxial rate of rotation (in rad/s) and degrees of rotation (in Degrees) from the mobile phone gyroscope (30 Hz)
- Triaxial geomagnetic field strength (in μT) from the mobile phone magnetometer (30 Hz)
- Latitude and longitude, and horizontal accuracy* (in meters) from the mobile phone GPS (every 5 seconds)
- Blood volume pressure (in nano Watt) from E4 photoplethysmography (PPG) sensor (64 Hz)
- Electrodermal activity (skin conductance in μS) from E4 EDA sensor (4 Hz)
- Average** heart rate values (in bps) computed in 10 seconds-span based on the BVP analysis from E4 (1 Hz)
- Peripheral skin temperature (in Celsius degrees) from E4 infrared thermopile (4 Hz)
* The estimated horizontal accuracy is defined as the radius of 68% confidence according to the API.
Reference: https://developer.android.com/reference/android/location/Location#getAccuracy()
** HR values are not derived from a real-time reading but are created after the data collection session.
Reference: https://support.empatica.com/hc/en-us/articles/360029469772-E4-data-HR-csv-explanation
레이블 파일은 12개의 열로 구성되어 있으며, 다음의 정보를 포함합니다.
Column name | Options (Descriptions)
|
ts
| timestamp (Unix epoch time)
|
action
| sleep, personal_care, work, study, household, care_housemem (caregiving), recreation_media, entertainment, outdoor_act (sports), hobby, recreation_etc (free time), shop, communitiy_interaction (regular activity), travel (includes commute), meal (includes snack), socialising |
actionOption
| Details of the selected action. See the description below.
|
actionSub
| meal_amount when action=meal or snack move_method when action=travel |
actionSubOption
| 1 (light), 2 (moderate), 3 (heavy) when actionSub=meal_amount 1 (walk), 2 (driving), 3 (taxi, passenger), 4 (personal mobility), 5 (bus), 6 (train, subway), 7 (others) when actionSub=move_method |
condition
| ALONE, WITH_ONE, WITH_MANY
|
conditionSub1Option
| 1 (with families), 2 (with friends), 3 (with colleagues), 4 (acquaintances), 5 (others)
|
conditionSub2Option
| 1 (passive in conversation), 2 (moderate participation in conversation), 3 (active in conversation)
|
place
| home, workplace, restaurant, outdoor, other_indoor
|
emotionPositive
| (negative) 1-2-3-4-5-6-7 (positive)
|
emotionTension
| (relaxed) 1-2-3-4-5-6-7 (aroused)
|
activity***
| 0 (IN_VEHICLE), 1 (ON_BICYCLE), 2 (ON_FOOT), 3 (STILL), 4 (UNKNOWN), 5 (TILTING), 7 (WALKING), 8 (RUNNING) |
*** Values in the activity column represent the detected activity of the mobile device using Google's Awareness API.
Reference: https://developers.google.com/android/reference/com/google/android/gms/location/DetectedActivity?hl=en
레이블 파일의 actionOption 열은 아래의 정보로 구성되어 있습니다.
==================================================================================================
111 Sleep
112 Sleepless
121 Meal
122 Snack
131 Medical services, treatments, sick rest
132 Personal hygiene (bath)
133 Appearance management (makeup, change of clothes)
134 Beauty-related services
211 Main job
212 Side job
213 Rest during work
22 Job search
311 School class / seminar (listening)
312 Break between classes
313 School homework, self-study (individual)
314 Team project (in groups)
321 Private tutoring (offline)
322 Online courses
41 Preparing food and washing dishes
42 Laundry and ironing
43 Housing management and cleaning
44 Vehicle management
45 Pet and plant caring
46 Purchasing goods and services (grocery/take-out)
51 Caring for children under 10 who live together
52 Caring for elementary, middle, and high school students over 10 who live together
53 Caring for a spouse
54 Caring for parents and grandparents who live together
55 Caring for other family members who live together
56 Caring for parents and grandparents who do not live together
57 Caring for other family members who do not live together
81 Personal care-related travel
82 Commuting and work-related travel
83 Education-related travel
84 Travel related to housing management
85 Travel related to caring for family and household members
86 Travel related to participation and volunteering
87 Socializing and leisure-related travel
61 Religious activities
62 Political activity
63 Ceremonial activities
64 Volunteer
711 Offline communication
712 Video or voice call
713 Text or email (Online)
721 Reading books, newspapers, and magazines
722 Watching TV or video
723 Listening to audio
724 Internet search or blogging
725 Gaming (mobile, computer, video)
741 Watching a sporting event
742 Watching movie
743 Concerts and plays
744 Art galleries and museums
744 Amusement Park, zoo
745 Festival, carnival
746 Driving, sightseeing, excursion
751 Walking
752 Running, jogging
753 Climbing, hiking
754 Biking
755 Ball games (soccer, basketball, baseball, tennis, etc)
756 Personal exercises (yoga, pilates, etc.)
756 Camping, fishing
761 Group games (board games, card games, puzzles, etc.)
762 Personal hobbies (woodworking, gardening, etc.)
763 Group performances (orchestra, choir, troupe, etc.)
764 Liberal arts and learning (languages, musical instruments, etc.)
791 Nightlife
792 Smoking
793 Do nothing and rest
91 Online shopping
92 Offline shopping
==================================================================================================
2020 데이터셋의 실험자별 데이터 통계는 다음과 같습니다.
******* 2020 데이터 업데이트: 2022. 3. 15 *******************************************************
1. 실험자의 인구통계학적 정보가 추가되었습니다. (성별, 나이, 주사용 손, 키, 몸무게)
2. 매일 오전 실험 시작 시 (AM), 그리고 실험 종료 시 (PM) 입력한 수면 관련 설문조사의 결과가 추가되었습니다.
Column | Descriptions |
amPm | am: Sleep related questionnaire was carried out in the morning. pm: Questionnaire collected factors that may have affect on sleep, performed in the evening. |
sleep | Q. Are you satisfied with your sleep? (Subjective sleep quality score) A. 1 (Not at all), 2 (Not much), 3 (Moderately)), 4 (Fairly), 5 (Fully)
|
sleepProblem | Q. Did you have problems during sleep? A. 1 (It took more than 30 minutes to fall asleep.), 2 (I was awake during the night or prior to my scheduled wake time.), 3 (I was awake during the night to go to the bathroom.), 4 (I snored loudly during the sleep or woke up during the night choking.), 5 (I was disturbed by the low temperature during sleep.), 6 (I was disturbed by the high temperature during sleep.), 7 (I had nightmares.), 8 (I was disturbed by the pain.), 9 (I was disturbed by other reasons not listed above.), 0 (I did not have any problems.)
|
dream
| 1 (Nightmare), 2 (Neutral dream), 3 (Nice dream), 4 (None)
|
amCondition
| Q. Do you feel refreshed after awakening? (Physical condition) A. 1 (Not at all), 2 (Not much), 3 (Moderately), 4 (Fairly), 5 (Fully)
|
amEmotion
| Q. How do you feel after awakening? (Emotional condition) A. 1 (Very unpleasant), 2 (Unpleasant), 3 (Moderate), 4 (Pleasant), 5 (Very pleasant)
|
pmEmotion
| Q. How do you feel now? (Emotional condition before you sleep) A. 1 (Very unpleasant), 2 (Unpleasant), 3 (Moderate), 4 (Pleasant), 5 (Very pleasant) |
pmStress
| Q. How stressed are you today? A. 1 (Very much), 2 (Fairly), 3 (As usual), 4 (Not much), 5 (Not at all)
|
pmFatigue
| Q. How tired are you today? (Physical condition) A. 1 (Very much), 2 (Fairly), 3 (As usual), 4 (Not much), 5 (Not at all) |
caffeine
| Types of beverages that contains caffeine, if any. |
cAmount
| Amount of caffeinated beverages (in ml).
|
alcohol
| Types of beverages that contains alcohol, if any.
|
aAmount | Amount of alcoholic beverages (in ml).
|
3. Withings Sleep Tracking Mat로 수집한 수면 측정 데이터가 추가되었습니다. (API)
실험 시 지급된 수면 측정 센서를 매트리스 아래에 설치하고 평소와 같이 수면을 취하도록 안내하였고,
수면 데이터는 Withings 모바일 앱을 통해 자동으로 서버로 전송되는 방식으로 수집하였습니다.
Column | Descriptions |
startDt | Start date. |
endDt | End date. |
lastUpdate | Timestamp for requesting data that were updated or created after this date. Useful for data synchronization between systems.
|
wakeupduration
| Time spent awake (in seconds).
|
lightsleepduration
| Duration in state light sleep (in seconds).
|
deepsleepduration
| Duration in state deep sleep (in seconds).
|
wakeupcount
| Number of times the user woke up while in bed. Does not include the number of times the user got out of bed.
|
durationtosleep
| Time to sleep (in seconds).
|
remsleepduration
| Duration in state REM sleep (in seconds).
|
durationtowakeup
| Time to wake up (in seconds).
|
hr_average
| Average heart rate.
|
hr_min
| Minimal heart rate.
|
hr_max
| Maximal heart rate.
|
rr_average
| Average respiration rate.
|
rr_min
| Minimal respiration rate.
|
rr_max
| Maximal respiration rate.
|
breathing_disturbances_intensity
| Wellness metric, available for all Sleep and Sleep Analyzer devices. Intensity of breathing disturbances
|
snoring
| Total snoring time
|
snoringepisodecount
| Numbers of snoring episodes of at least one minute
|
sleep_score
| Sleep score
|
*******************************************************************************************************
*******************************************************************************************************
* ETRI_Lifelog_Dataset_2019_2018
2019, 2018 데이터셋은 다음의 데이터를 포함합니다.
==================================================================================================
- Triaxial acceleration force (in m/s^2) from the mobile phone accelerometer (30 Hz) and E4 accelerometer (32 Hz)
- Triaxial rate of rotation (in rad/s) and degrees of rotation (in Degrees) from the mobile phone gyroscope (30 Hz)
- Triaxial geomagnetic field strength (in μT) from the mobile phone magnetometer (30 Hz)
- Latitude and longitude, and horizontal accuracy* (in meters) from the mobile phone GPS (every 1 minute)
- Blood volume pressure (in nano Watt) from E4 photoplethysmography (PPG) sensor (64 Hz)
- Electrodermal activity (skin conductance in μS) from E4 EDA sensor (4 Hz)
- Average** heart rate values (in bps) computed in 10 seconds-span based on the BVP analysis from E4 (1 Hz)
- Peripheral skin temperature (in Celsius degrees) from E4 infrared thermopile (4 Hz)
* The estimated horizontal accuracy is defined as the radius of 68% confidence according to the API.
Reference: https://developer.android.com/reference/android/location/Location#getAccuracy()
** HR values are not derived from a real-time reading but are created after the data collection session.
Reference: https://support.empatica.com/hc/en-us/articles/360029469772-E4-data-HR-csv-explanation
레이블 파일은 12개의 열로 구성되어 있으며, 다음의 정보를 포함합니다.
Column name | Options (Descriptions)
|
ts
| timestamp (Unix epoch time)
|
action
| sleep, personal_care, work, study, household, recreation_media, entertainment, outdoor_act (sports), hobby, recreation_etc (free time), communitiy_interaction (regular activity), travel (includes commute), meal (includes snack), socialising |
actionOption
| 0 (walking upstairs), 1 (walking downstairs), 2 (walking), 3 (sitting), 4 (lying), 5 (standing), 6 (running)
|
actionSub | 0 (on the table), 1 (in the pocket), 2 (in hand), 3 (others), 4 (in the bag) |
actionSubOption
| 0 (on bicycle), 1 (on foot), 2 (driving), 3 (public transportation), 4 (personal mobility)
|
condition
| 0 (alone), 1 (family member), 2 (friend), 3 (collegue), 4 (acquaintance)
|
conditionSub1Option
| 1 (family member), 2 (friend), 3 (collegue), 4 (acquaintance)
|
conditionSub2Option
| 1 (online)
|
place
| home, workplace, restaurant, outdoor, other_indoor
|
emotionPositive
| (negative) 1-2-3-4-5-6-7 (positive)
|
emotionTension
| (relaxed) 1-2-3-4-5-6-7 (aroused)
|
activity***
| 0 (IN_VEHICLE), 1 (ON_BICYCLE), 2 (ON_FOOT), 3 (STILL), 4 (UNKNOWN), 5 (TILTING), 7 (WALKING), 8 (RUNNING) |
*** Values in the activity column represent the detected activity of the mobile device using Google's Awareness API.
Reference: https://developers.google.com/android/reference/com/google/android/gms/location/DetectedActivity?hl=en
2019, 2018 데이터셋의 실험자별 데이터 통계는 다음과 같습니다.
******* 2019-2018 데이터 업데이트: 2022. 3. 17 ************************************************
1. 실험자의 인구통계학적 정보가 추가되었습니다. (성별, 나이, 키, 몸무게)
2. 매일 오전 실험 시작 시 (AM), 그리고 실험 종료 시 (PM) 입력한 수면 관련 설문조사의 결과가 추가되었습니다.
Column | Descriptions |
amPm | am: Sleep related questionnaire was carried out in the morning. pm: Questionnaire collected factors that may have affect on sleep, performed in the evening. |
sleep | Q. Are you satisfied with your sleep? (Subjective sleep quality score) A. 1 (Not at all), 2 (Not much), 3 (Moderately)), 4 (Fairly), 5 (Fully)
|
sleepProblem | Q. Did you have problems during sleep? A. 1 (It took more than 30 minutes to fall asleep.), 2 (I was awake during the night or prior to my scheduled wake time.), 3 (I was awake during the night to go to the bathroom.), 4 (I snored loudly during the sleep or woke up during the night choking.), 5 (I was disturbed by the low temperature during sleep.), 6 (I was disturbed by the high temperature during sleep.), 7 (I had nightmares.), 8 (I was disturbed by the pain.), 9 (I was disturbed by other reasons not listed above.), 0 (I did not have any problems.)
|
dream
| 1 (Nightmare), 2 (Neutral dream), 3 (Nice dream), 4 (None)
|
amCondition
| Q. Do you feel refreshed after awakening? (Physical condition) A. 1 (Not at all), 2 (Not much), 3 (Moderately), 4 (Fairly), 5 (Fully)
|
amEmotion
| Q. How do you feel after awakening? (Emotional condition) A. 1 (Very unpleasant), 2 (Unpleasant), 3 (Moderate), 4 (Pleasant), 5 (Very pleasant)
|
pmEmotion
| Q. How do you feel now? (Emotional condition before you sleep) A. 1 (Very unpleasant), 2 (Unpleasant), 3 (Moderate), 4 (Pleasant), 5 (Very pleasant) |
pmStress
| Q. How stressed are you today? A. 1 (Very much), 2 (Fairly), 3 (As usual), 4 (Not much), 5 (Not at all)
|
ifUnusual
| Q. Was it an unusual day? A. 1 (Yes), 2 (No) |
breakfast
| Q. Were you satisfied with today's breakfast? A. 1 (Not at all), 2 (Moderately), 3 (Fully), 4 (Not applicable) |
lunch
| Q. Were you satisfied with today's lunch? A. 1 (Not at all), 2 (Moderately), 3 (Fully), 4 (Not applicable)
|
dinner
| Q. Were you satisfied with today's dinner? A. 1 (Not at all), 2 (Moderately), 3 (Fully), 4 (Not applicable)
|
lateSnack
| Q. Were you satisfied with today's midnight snack? A. 1 (Not at all), 2 (Moderately), 3 (Fully), 4 (Not applicable)
|
amCaffeine
| Types of caffeinated beverages consumed in the morning.
|
amCaffAmount
| Amount of caffeinated beverages consumed in the morning (in cups).
|
pmCaffeine
| Types of caffeinated beverages consumed in the afternoon.
|
pmCaffAmount
| Amount of caffeinated beverages consumed in the afternoon (in cups).
|
alcohol
| Types of beverages that contains alcohol, if any.
|
aAmount | Amount of alcoholic beverages (in bottles).
|
3. ActiGraph (2019 dataset), Fitbit Versa (2018 dataset)로 수집한 수면 측정 데이터가 추가되었습니다.
Column | Descriptions |
startDt | Start date. |
endDt | End date. |
device | Device used to measure the sleep related data during the night sleep. Actigraph (2019), Fitbit Versa (2018) |
sleep_score
| Sleep score.
|
total_sleep_time
| Duration of sleep (in seconds).
|
time_in_bed
| Duration of time in bed (in seconds).
|
waso
| Wake After Sleep Onset (WASO)
|
wakeupcount
| Number of times the user woke up while in bed. Does not include the number of times the user got out of bed.
|
aal
| Average Awakening Length (AAL)
|
movement_index
| Movement Index (MI), The total of scored awake minutes divided by Total time in bed in hours x 100.
|
fragmentation_index
| Fragmentation Index (FI),The percentage of one minute periods of sleep vs. all periods of sleep in the sleep period.
|
sleep_fragmentation_index
| The sum of MI and FI.
|
*******************************************************************************************************
*******************************************************************************************************
* 본 데이터셋을 활용할 경우 다음 논문을 인용 바랍니다.
Seungeun Chung, Chi Yoon Jeong, Jeong Mook Lim, Jiyoun Lim, Kyoung Ju Noh, Gague Kim, Hyuntae Jeong,
Real-world multimodal lifelog dataset for human behavior study. ETRI Journal 43(6), 2021
https://doi.org/10.4218/etrij.2020-0446
* 관련 논문
Jiyoun lim, Chi Yoon Jeong, Jeong Muk Lim, Seungeun Chung, Gague Kim, Kyoung Ju Noh, Hyuntae Jeong
Assessing Sleep Quality Using Mobile EMAs: Opportunities, Practical Consideration, and Challenges. IEEE Access 10, 2022
https://doi.org/10.1109/ACCESS.2021.3140074
Seungeun Chung, Jiyoun Lim, Kyoung Ju Noh, Gague Kim, Hyuntae Jeong,
Sensor Data Acquisition and Multimodal Sensor Fusion for Human Activity Recognition Using Deep Learning. Sensors 19(7), 2019
https://doi.org/10.3390/s19071716
Jiyoun Lim, Seunghee Yoo, Seungeun Chung, Gague Kim, Kyoung Ju Noh, Jeong Muk Lim, Hyuntae Jeong:
SPER: Stay-Point Extraction considering Revisits in a Single Trajectory. ICTC 2021
https://doi.org/10.1109/ICTC52510.2021.9621139
임호연, 정승은, 정치윤, 정현태,
라이프로그 기반 일상생활 활동유형에 대한 탐색적 연구, 한국정보처리학회 추계 학술대회, 2020
https://library.etri.re.kr/service/rsch/etri-article/down.htm?view=open&resultId=0000062797
Jiyoun Lim, Seungeun Chung, Kyoung Ju Noh, Gague Kim, Hyun-Tae Jeong,
An empirical study on finding experience sampling parameters to explain sleep quality based on dimension reduction. ICTC 2019
https://doi.org/10.1109/ICTC46691.2019.8939976
Seungeun Chung, Inyoung Hwang, Jiyoun Lim, Hyun-Tae Jeong,
Finding Points-of-Interest (PoIs) from Life-logging and Location Trace Data. ICTC 2019
https://doi.org/10.1109/ICTC46691.2019.8940021
Seungeun Chung, Jiyoun Lim, Kyoung Ju Noh, Gague Kim, Hyun-Tae Jeong,
Sensor Positioning and Data Acquisition for Activity Recognition using Deep Learning. ICTC 2018
https://doi.org/10.1109/ICTC.2018.8539473