What are the overarching themes for our analyses?

We are currently interested in developing the following types of analysis that can be applied to individualized data.  As you can see below, each approach is in various stages of development, and more input is needed before we can reach a consensus on the best approaches for each situation.  The criteria for applying an analytical framework are listed under 'Criteria.'

Intervention Analysis

One of the most basic questions for which Big Data can be usefully applied is to answer the question, 'Did intervention X make a difference?'  In this section, we describe methods to analyze both continuous and binary longitudinal outcomes for which at some point in time a change was made to a covariate, to answer the question of whether that change made a statistically significant difference.

CORRELATIONS

When collecting two variables over time on any individual, the first question that should be answered is 'Are they correlated?' In this section, we describe methods that can be applied between two longitudinal variables to determine whether any correlation exists between these two, either at the same time, or with one lagging or leading the other.  

SEASONAL PATTERNS

Nature is cyclical, and in order to understand data about ourselves it is important that we develop methods to describe the seasonality of our data.  By 'seasons' we can mean the cycle that occurs over the course of a year, or over a single day.  Understanding the seasonality of our data is critical to avoiding bias in our findings, and can also provide insight into lifestyle or behavioral factors that can be contributing to the patterns we see.