newcastleuniversity / dispel

A python DIgital Signal ProcEssing Library developed to standardize extraction of sensor-derived measures (SDMs) from wearables or smartphones data.

---

DISPEL is a python DIgital Signal ProcEssing Library developed to standardize extraction of sensor-derived measures (SDMs) from wearables or smartphones data.

• 👋 Goodbye to Guesswork: Transparency is key, and with DISPEL, you can track and standardize every step of your data analysis, ditching the guesswork for good.

• 💸 Kick Clinical Trial Costs to the Curb: Say hello to DISPEL, your new best friend for slashing the hefty price tag and long timelines of clinical trials. Join the revolution !

• 🐍 Python Power: Tap into the power of Python with DISPEL. It's user-friendly, packed with helpful guides, and sprinkled with examples to get your data dancing to your tune.

• 🌊 Embrace the Data Deluge: Don't drown in data—master it. DISPEL devours data from any source, spits it out formatted, and readies it for some serious number crunching.

• 💻 Code That Cares: Our object-oriented codebase is the heart of a centralized hub designed for all your data analysis needs, making your workflow smooth and seamless.

• 🔬 Customize Like a Pro: Flex your research muscles with DISPEL's modular pipeline, which adapts to your project's needs, skips the snags, and keeps the insights rolling in.

• 🏆 Real-World Ready: Step out of the lab and into the world with confidence. DISPEL equips you with the tools to factor in real-world variables, ensuring your research stays relevant and resilient.

Contribution

Join our efforts in enhancing DISPEL's capabilities by contributing new processing modules and improvements; for detailed guidelines, please refer to our CONTRIBUTING file.

Documentation

See our documentation for how to install and get started with DISPEL.

The Team

DISPEL is a community-driven project with several skillful engineers and researchers contributing to it.

DISPEL is currently maintained by Chloe Hinchliffe, Alf Scotland, Adrien Juraver, Angelos Karatsidis and Gautier Cosne, with major contributions of talented individuals in various forms and means. A non-exhaustive list needs to mention: Joaquin Penalver de Andres, Ali Neishabouri, Clément Dulong, Oussama Tchita, Loïc Carment, Angéline Plaud, Kevin Bouaou.

License

DISPEL has a MIT-license, as found in the LICENSE file.

dispel's Issues

Add descriptions of test-specific assumptions

Overview

Currently, little information is provided on the basic assumptions for each algorithm in DISPEL.
For instance, gait and turn algorithms assume a lumbar-mounted device. Drawing algorithms assume a handheld device.

Objectives

1. Add information about BDH/ADS Konectom battery of tests based this link
2. Add information about APDM / Mobilise-D data based on this link
3. Add information about Digital Artefacts Finger Tapping this link

Desired Outcomes

• More information about test-specific assumptions and instructions added as module doctstrings.

Create a Tutorial Notebook to leverage DISPEL with Mobilize-D YAR Data

Overview

This issue aims to develop a comprehensive, easy-to-follow tutorial notebook that effectively demonstrates how to leverage DISPEL with Mobilize-D data. This tutorial aims to provide new users with a hands-on approach to learning the library's functionalities, making it easier to integrate into their projects.

Objectives

1. Data Introduction: Briefly introduce Mobilize-D data, its significance, and typical use cases.
2. Environment Setup: Guide on setting up the necessary environment, including library installation.
3. Basic Operations: Cover basic library functions, showcasing simple data manipulations and operations.
4. Advanced Features: Dive into more complex features, providing real-world examples deriving SDMs from Mobilize-D data.
5. Best Practices: Offer guidance on best practices for transforming datasets and including technical and behavioural deviations.
6. Interactive Examples: Include interactive examples for users to experiment with, enhancing learning engagement.

Desired Outcomes

• A Jupyter notebook that is well-documented, with clear explanations and code comments.
• The tutorial should cater to beginners and intermediate users, gradually building in complexity.
• Users should feel confident using the library with Mobilize-D data for their specific needs after completion.