Skip to content

Implementation of a Sensor Data Analysis Pipeline

This sensor data analysis pipeline is designed to benefit all members of DSI Health conducting studies with sensors. Inspired by the observed overlaps in data analysis workflows across various specialties within our community, this project aims to create a comprehensive and versatile library. The design, development, and evaluation of this library are conducted collaboratively with interested DSI Health members, ensuring its relevance and utility.

Motivation

  • Many of us at the DSI work on sensor data projects, either hiring someone for this analysis, or coding the analysis ourselves
  • Much of that sensor data needs pre-processing and visualizing in similar ways
  • As a result, much of the workflow/code is repeated throughout our community
  • Thus, this work should be made towards a centrally accessible tool within the DSI

Aims

  • A tool that supports us all in our data-driven decision making, training machine learning models, finding data relationships, etc.— so we don’t rewrite the same data analysis code over and over again
  • Building upon the findings from the DSI Health project from 2023
  • Collecting and characterizing use cases for sensor data analysis within the DSI
  • Identifying similarities in analysis and visualization steps such analyses
  • Mapping out a generalizable workflow for our community
  • Open sourced, and centrally accessible tool within the DSI

Call for participation

Timeline

Jan-Feb

 

PLANNING:

  • Review of 2023 project data
  • Collect community cases
  • Priority setting
  • Structuring subgroups
  • DSI Infrastructure Call application

Apr

MAIN WORKSHOP 1: 

  • Launch
  • Review case studies
  • Identify building blocks
  • Milestone planning
  • Subgroup organization

May-Sept

SUBGROUP WORKSHOPS:

2-3 rounds of iterative work on identified building blocks within the subgroups

Oct

 

 

Nov- Dec

MAIN WORKSHOP 2: 

Progress review

 

Reflection and preparation for future steps