Figure 3 shows the relationship between the composite well-being indices and the
obtained EQ5 scores of 107 respondents, 3 of which filled the questionnaire three
times, 20 filled it twice, and the rest once during the observed data collection period.
The analysis reveals a medium positive correlation of 0.424 (with the p-value of
approx. 3.7 Â 10
−7
) between our constructed composite index and cumulative EQ5
scores. The composite indicator and its constituting domain indices can therefore be
considered promising for their intended purpose of basic represention of the urban
physical well-being aspects modelled from a variety of heterogeneous underlying
activity and health/vital parameters measured by IoT wearable devices, summarized in
4 main dimensional and one overall derived score convenient for comparisons, interpretation and presentation to the end-users, particularly in the required shortest most
concise manner and form, such as through a mobile app. UI or intervention messages.
As almost half of the questions in EQ5 are very remotely or not at all related to the
physical well-being aspects summarized by the composite indicator, the correlation is
expected to increase when the ongoing work in incorporating social and other wellbeung aspects fully in the model is completed.
Found small positive correlation of 0.287 (p-value 0.001) between the composite
indicator and cumulative scores of ESS questionnaire (in which most of the questions
are not related to physical well-being) additionally points to the significance of this
composite indicator to the overall well-being. The work also continues on the cleaning
and pre-processing the data collected on the remaining monitored citizens, incorporation of machine learning methods in the model and exploration and modelling of the
influence of detailed sleep and cardiac parameters, as well as of the sensed ambiental
data, on the main well-being domains.
Fig. 3. Scatter plot of composite well-being index vs. cumulative EQ5 score values
Baseline Modelling and Composite Representation
165
obtained EQ5 scores of 107 respondents, 3 of which filled the questionnaire three
times, 20 filled it twice, and the rest once during the observed data collection period.
The analysis reveals a medium positive correlation of 0.424 (with the p-value of
approx. 3.7 Â 10
−7
) between our constructed composite index and cumulative EQ5
scores. The composite indicator and its constituting domain indices can therefore be
considered promising for their intended purpose of basic represention of the urban
physical well-being aspects modelled from a variety of heterogeneous underlying
activity and health/vital parameters measured by IoT wearable devices, summarized in
4 main dimensional and one overall derived score convenient for comparisons, interpretation and presentation to the end-users, particularly in the required shortest most
concise manner and form, such as through a mobile app. UI or intervention messages.
As almost half of the questions in EQ5 are very remotely or not at all related to the
physical well-being aspects summarized by the composite indicator, the correlation is
expected to increase when the ongoing work in incorporating social and other wellbeung aspects fully in the model is completed.
Found small positive correlation of 0.287 (p-value 0.001) between the composite
indicator and cumulative scores of ESS questionnaire (in which most of the questions
are not related to physical well-being) additionally points to the significance of this
composite indicator to the overall well-being. The work also continues on the cleaning
and pre-processing the data collected on the remaining monitored citizens, incorporation of machine learning methods in the model and exploration and modelling of the
influence of detailed sleep and cardiac parameters, as well as of the sensed ambiental
data, on the main well-being domains.
Fig. 3. Scatter plot of composite well-being index vs. cumulative EQ5 score values
Baseline Modelling and Composite Representation
165