EXCLUDE_SLEEP module for steps intraday featuresLuca Canzian and Mirco Musolesi. 2015. Trajectories of depression: unobtrusive monitoring of depressive states by means of smartphone mobility traces analysis. In Proceedings of the 2015 ACM International Joint Conference on Pervasive and Ubiquitous Computing (UbiComp ‘15). Association for Computing Machinery, New York, NY, USA, 1293–1304. DOI:https://doi.org/10.1145/2750858.2805845
If you computed locations features using the provider [PHONE_LOCATIONS][DORYAB] cite this paper and this paper in addition to RAPIDS.
If you computed locations features using the provider [PHONE_LOCATIONS][DORYAB] cite this paper and this paper in addition to RAPIDS. In addition, if you used the SUN_LI_VEGA_STRATEGY strategy, cite this paper as well.
Doryab et al. citation
Doryab, A., Chikarsel, P., Liu, X., & Dey, A. K. (2019). Extraction of Behavioral Features from Smartphone and Wearable Data. ArXiv:1812.10394 [Cs, Stat]. http://arxiv.org/abs/1812.10394
@@ -1885,6 +1885,10 @@ URL: https://preprints.jmir.Canzian et al. citation
Luca Canzian and Mirco Musolesi. 2015. Trajectories of depression: unobtrusive monitoring of depressive states by means of smartphone mobility traces analysis. In Proceedings of the 2015 ACM International Joint Conference on Pervasive and Ubiquitous Computing (UbiComp ‘15). Association for Computing Machinery, New York, NY, USA, 1293–1304. DOI:https://doi.org/10.1145/2750858.2805845
Sun et al. citation
+Sun S, Folarin AA, Ranjan Y, Rashid Z, Conde P, Stewart C, Cummins N, Matcham F, Dalla Costa G, Simblett S, Leocani L, Lamers F, Sørensen PS, Buron M, Zabalza A, Guerrero Pérez AI, Penninx BW, Siddi S, Haro JM, Myin-Germeys I, Rintala A, Wykes T, Narayan VA, Comi G, Hotopf M, Dobson RJ, RADAR-CNS Consortium. Using Smartphones and Wearable Devices to Monitor Behavioral Changes During COVID-19. J Med Internet Res 2020;22(9):e19992
+Description
+- data/raw/{pid}/phone_locations_raw.csv
- data/interim/{pid}/phone_locations_processed.csv
- data/interim/{pid}/phone_locations_processed_with_datetime.csv
-- data/interim/{pid}/phone_locations_processed_with_datetime_with_home.csv
+- data/interim/{pid}/phone_locations_processed_with_datetime_with_doryab_columns.csv
- data/interim/{pid}/phone_locations_features/phone_locations_{language}_{provider_key}.csv
- data/processed/features/{pid}/phone_locations.csv
Fine-Tuning Clustering Parameters
Based on an experiment where we collected fused location data for 7 days with a mean accuracy of 86 & SD of 350.874635, we determined that EPS/MAX_EPS=100 produced closer clustering results to reality. Higher values (>100) missed out on some significant places, like a short grocery visit, while lower values (<100) picked up traffic lights and stop signs while driving as significant locations. We recommend you set EPS based on your location data’s accuracy (the more accurate your data is, the lower you should be able to set EPS).
Duration Calculation -To calculate the time duration component for our features, we compute the difference between consecutive rows’ timestamps to take into account sampling rate variability. If this time difference is larger than a threshold (300 seconds by default), we replace it with a maximum duration (60 seconds by default, i.e., we assume a participant spent at least 60 seconds in their last known location)
-Home location
-Home is calculated using all location data of a participant between 12 am and 6 am, then applying a clustering algorithm (DB_SCAN or OPTICS) and considering the center of the biggest cluster home for that participant.
Home location
+DORYAB_STRATEGY: home is calculated using all location data of a participant between 12 am and 6 am, then applying a clustering algorithm (DBSCAN or OPTICS) and considering the center of the biggest cluster home for that participant.
SUN_LI_VEGA_STRATEGY: home is calculated using all location data of a participant between 12 am and 6 am, then applying a clustering algorithm (DBSCAN or OPTICS). The following steps are used to infer the home location per day for that participant:
if there are records within [03:30:00, 04:30:00] for that night:
+ we choose the most common cluster during that period as a home candidate for that day.
+ elif there are records within [midnight, 03:30:00) for that night:
+ we choose the last valid cluster during that period as a home candidate for that day.
+ elif there are records within (04:30:00, 06:00:00] for that night:
+ we choose the first valid cluster during that period as a home candidate for that day.
+ else:
+ the home location is NA (missing) for that day.
If the count of consecutive days with the same candidate home location cluster label is larger or equal to [MINIMUM_DAYS_TO_DETECT_HOME_CHANGES],
+ the candidate will be regarded as the home cluster; otherwise, the home cluster will be the last valid day’s cluster.
+ If there are no valid clusters before that day, the first home location in the days after is used.