Pattern discovery in continuous glucose monitoring data

A method comprising:

• acquiring multiple data series comprised of measurements generated by a glucose monitoring device,
• wherein each data series of the multiple data series corresponds to a different multi-hour interval of time at least four hours in length over which a blood glucose level of a subject is sampled by the glucose monitoring device at a frequency of at least every ten minutes;
• acquiring a target data series comprised of fabricated measurements that mimic a time-varying blood glucose level of an individual without diabetes;
• for each data series of the multiple data series,
• producing a set of distance measures by (i) producing a distance measure for the target data series and (ii) producing a distance measure for each other data series of the multiple data series, so as to produce multiple sets of distance measures for the multiple data series,
• wherein each distance measure in the set of distance measures is indicative of similarity between a corresponding pair of data series of the multiple data series, and
• wherein each set of data series of the multiple sets of distance measures is associated with a different one of the multiple data series;
• populating the multiple sets of distance measures produced for the multiple data series into a matrix having an equal number of rows and columns;
• sorting the multiple data series into one or more clusters based on an analysis of the multiple sets of distance measures in the matrix;
• causing display of (i) a visualization in which a first cluster of the one or more clusters that includes the target data series is visually distinguishable from other clusters, if any, of the one or more clusters and (ii) information gleaned from an analysis of the one or more clusters first cluster on a first interface that is accessible, via a first computing device, to a person responsible for monitoring glycemic health of the subject;
• receiving input that is indicative of feedback that is provided by the person through the first interface and that is personalized for the subject based on the information;
• and
• causing display of content that includes, or is based on, the feedback on a second interface that is accessible, via a second computing device, to the subject.

A non-transitory computer-readable medium with instructions stored thereon that, when executed by a processor, cause the processor to perform operations comprising:

• acquiring multiple data series associated with a subject,
• wherein each data series of the multiple data series corresponds to a different multi-hour interval of time over which a blood glucose level of the subject is sampled sufficiently frequently to detect variations due to foodstuff consumption;
• acquiring a target data series comprised of fabricated measurements that mimic a blood glucose level of an individual without diabetes;
• for each data series of the multiple data series,
• producing a set of distance measures by (i) producing a distance measure for the target data series and (ii) producing a distance measure for each other data series of the multiple data series, so as to produce multiple sets of distance measures for the multiple data series,
• wherein each distance measure in the set of distance measures is indicative of similarity between a corresponding pair of data series of the multiple data series, and
• wherein each set of data series of the multiple sets of distance measures is associated with a different one of the multiple data series;
• and
• creating a similarity matrix by populating a data structure with the multiple sets of distance measures produced for the multiple data series;
• causing display of a visualization in which the target data series and the multiple data series are represented as traces with respect to a common timeline on a first interface that is accessible, via a first computing device, to a person responsible for monitoring glycemic health of the subject;
• receiving, through the first interface, first input indicative of a selection of a first trace that corresponds to a first data series:
• causing the visualization to be adjusted such that a second trace corresponding to a second data series, determined to be most similar to the first data series based on an analysis of the similarity matrix, is made visually identifiable;
• receiving, through the first interface, second input that is indicative of feedback that is provided by the person;
• and
• causing display of content that includes, or is based on, the feedback on a second interface that is accessible, via a second computing device, to the subject.