F R M e W O R K
Dutta et al. (2019)

Complaint Analytics

The complaint analytics function calculates the frequency of hot/cold-related occupant complaints and the conditions that produce them.

Outputs

Figures

The complaint analytics function generates three visualizations intended to depict the distribution of complaints, illustrate the relationship between outdoor and indoor air conditions and complaints, and the prevailing conditions which give rise to complaints. The first visual categorizes the complaints by the type of complaint (hot or cold related) and counts the number of complaints by the month and period of the day that the complaint was entered. Figure 1 is an example of this visual.

A generic square placeholder image with rounded corners in a figure.
Figure 1: Example visual of distribution of complaints by type of complaints, and month and time of day that they were entered.

The second visualization illustrates the relationship between hot and cold related complaints and the outdoor and indoor air temperatures at the time the complaint was entered. Figure 2 is an example of this visual.

A generic square placeholder image with rounded corners in a figure.
Figure 2: Example visual depicting complaints in relation to outdoor and indoor air temperature at the time they were entered.

The third visualization is a decision tree diagram that predicts the proportion of complaints that would be made with respect to certain conditions. These criteria for consideration are outdoor air temperature, time of day, and the day of the week. The boxes which branch to the left represent the predicted proportion of complaints when the condition in the preceding box is satisfied. For example, if a box with the condition, 'Hour of the day <= 10' has a left branch with 40% and a right branch with 60%, it is predicted that 40% of all complaints made will occur before 10 am. The condition is displayed at the top of the box, and the predicted proportion is displayed at the center of each box between the 0.0s. The 0.0s are only placeholders and can be ignored. Figure 3 is an example of this visual.

A generic square placeholder image with rounded corners in a figure.
Figure 3: Example visual of decision tree diagram that predicts the proportion of complaints that would be made with respect to outdoor air temperature, time of day, and the day of the week.

Key Performance Indicators (KPIs)

The complaint analytics function calculates the daily frequency of hot and cold complaints in the heating and cooling season separately. The values represent the number of complaints made per day. Higher frequencies indicate a higher rate of occurrence of a type of complaint for the particular season. Thus, a lower value is desirable. Table 1 is an example of the calculated frequencies.

Type of complaint Daily complaint frequency for heating season Daily complaint frequency for cooling season
Hot complaint 0.5 0
Cold complaint 0.2 0.8
Table 1: Example calculated daily frequencies of hot and cold complaints for the heating and cooling season.
Actively in development, an open-source project by the Data-driven Building Operation and Maintenance team within Carleton University's Building Performance Research Center.