Threshold Checker

This example shows how to use the functions apply_thresholds to check if a DMO falls within the thresholds for a specific cohort and height.

import pandas as pd
from mobgap import PROJECT_ROOT
from mobgap.aggregation import apply_thresholds, get_mobilised_dmo_thresholds

Selecting Example Data

We use some example DMO data. Note, that the apply_thresholds function can only be used with the data of a single participant at the time, as meta-data such as the participant’s height is required. Luckily, the all the example data is from the same participant, just recorded at different times.

DATA_PATH = PROJECT_ROOT / "example_data/original_results/mobilised_aggregator"

data = pd.read_csv(
    DATA_PATH / "aggregation_test_input.csv", index_col=0
).set_index(["visit_type", "participant_id", "measurement_date", "wb_id"])
data.head()

⤢

				duration_s	n_steps	cadence_spm	walking_speed_mps	stride_length_m	stride_duration_s	n_turns
visit_type	participant_id	measurement_date	wb_id
T1	12345	2023-01-01	0	4.74702	8	99.82188	1.16079	2.51885	1.58675	0
			1	5.13150	7	101.16429	2.57881	1.57243	1.46537	0
			2	8.52727	12	86.53527	1.60044	1.66305	2.56092	2
			3	16.24554	27	91.49977	0.95558	0.88961	3.14549	1
			4	6.09907	8	93.69895	2.33230	1.95969	2.35295	0

Load Thresholds

thresholds = get_mobilised_dmo_thresholds()
thresholds

⤢

	condition	free_living		laboratory		literature		global
	threshold_type	min	max	min	max	min	max	min	max
dmo	cohort
cadence_spm	CHF	40.594770	167.942930	53.617703	122.657158	36.20	150.900	40.000000	172.900
	COPD	38.880484	151.558718	59.451004	164.969764	71.20	154.800	40.000000	172.900
	HA	35.751898	156.946955	60.176644	146.309393	90.40	169.200	40.000000	172.900
	MS	38.443211	155.394496	44.016149	140.939597	45.60	164.400	40.000000	172.900
	PD	40.957969	142.121947	47.681576	147.054849	33.80	77.600	40.000000	172.900
	PFF	42.191719	157.613665	46.839404	133.129276	42.90	172.900	40.000000	172.900
walking_speed_mps	CHF	0.112478	1.757103	0.273401	1.528913	0.63	1.830	0.081515	2.220
	COPD	0.090731	1.641773	0.159689	1.918146	0.60	1.840	0.081515	2.220
	HA	0.097413	1.965728	0.207887	1.583779	0.78	2.220	0.081515	2.220
	MS	0.085918	1.920262	0.203213	1.877691	0.52	1.740	0.081515	2.220
	PD	0.081515	1.735348	0.139654	1.939697	0.56	2.100	0.081515	2.220
	PFF	0.104547	1.553186	0.143724	1.627383	0.11	1.590	0.081515	2.220
stride_length_m	CHF	0.185308	2.166646	0.444041	1.568583	0.53	1.860	0.150523	2.190
	COPD	0.176403	1.711645	0.223537	1.623733	1.03	2.190	0.150523	2.190
	HA	0.155126	2.024694	0.290283	1.669609	0.87	1.980	0.150523	2.190
	MS	0.191099	1.940537	0.363004	1.721322	0.72	2.100	0.150523	2.190
	PD	0.150523	1.982926	0.253069	1.772536	0.56	1.620	0.150523	2.190
	PFF	0.206787	1.697251	0.233460	1.587144	0.71	1.930	0.150523	2.190
stride_duration_s	CHF	0.702000	3.030857	0.980000	2.274545	0.46	1.320	0.460000	3.000
	COPD	0.770000	3.000000	0.727500	2.160000	0.68	1.410	0.460000	3.000
	HA	0.735435	3.254400	0.829091	2.142000	0.69	2.304	0.460000	3.000
	MS	0.775597	3.057750	0.855000	2.731667	0.69	2.304	0.460000	3.000
	PD	0.836253	2.928000	0.818571	2.522000	0.76	1.690	0.460000	3.000
	PFF	0.744000	2.769000	0.902727	2.584545	0.14	0.980	0.460000	3.000
step_duration_s	CHF	0.376000	1.504500	0.498000	1.144167	0.14	0.980	0.140000	2.124
	COPD	0.390400	1.788000	0.366000	1.181429	0.16	0.920	0.140000	2.124
	HA	0.367972	1.730400	0.410000	1.137500	0.25	1.040	0.140000	2.124
	MS	0.388084	1.905000	0.432857	1.360000	0.25	1.090	0.140000	2.124
	PD	0.417216	1.605600	0.415000	1.450000	0.19	1.100	0.140000	2.124
	PFF	0.371429	2.124000	0.430000	1.320000	0.35	1.100	0.140000	2.124

Apply Thresholds

data_mask = apply_thresholds(
    data,
    thresholds,
    cohort="CHF",
    height_m=1.75,
    measurement_condition="free_living",
)
data_mask.head()

⤢

				duration_s	n_steps	cadence_spm	walking_speed_mps	stride_length_m	stride_duration_s	n_turns
visit_type	participant_id	measurement_date	wb_id
T1	12345	2023-01-01	0	NaN	NaN	True	True	True	True	NaN
			1	NaN	NaN	True	False	True	True	NaN
			2	NaN	NaN	True	True	True	True	NaN
			3	NaN	NaN	True	True	True	False	NaN
			4	NaN	NaN	True	False	True	True	NaN

We can see that the output has exactly the same structure as the input data, but with boolean values indicating if the DMO falls within the thresholds or not. Columns for which no thresholds were provided are simply NaN, indicating that no filtering was applied. It is up to you how you want to use this information.

This output data can be used to filter the input data for DMOs that fall within the thresholds. This can be done in combination with the aggregation algorithm to only include DMOs that fall within the thresholds. See the mobilised_aggregator example for more information.

Estimated memory usage: 80 MB