Skip to content

pypi downloads versions discord

LABDA: Library for Advanced Behavioural Data Analysis

Introducing LABDA

LABDA, a Python package built for health researchers, transforms raw movement data into compelling stories about human behaviour.

The package is helpful for researchers who need to collect data on people's movement behaviour. LABDA acts as a translator for the language of movement, revealing not just the "how much" but also the fascinating story behind it. By analysing "when" people are active (morning jog vs evening stroll), "where" movement happens (gym workout vs park visit), and even the "why" behind activity choices, LABDA empowers researchers to paint a richer picture of physical activity patterns. This deeper understanding unlocks a universe of possibilities for exploring the connections between human movement, health and behaviour.

Simplified Workflows & Trustworthy Results

Data wrangling is rarely a perfect science, with endless debates on the "one true way." But sometimes, getting results trumps finding the absolute best method. That's where LABDA steps in! This user-friendly package prioritises transparency above all else. Even if it takes some shortcuts for efficiency, LABDA meticulously documents every processing step. This ensures you can trust your results and understand exactly how LABDA helped you get there. Let's explore the world of LABDA and see how it can streamline your data analysis journey!

LABDA Overview LABDA Overview
Fig. 1 LABDA Overview

Structure: The Struggle with Research Data

Have you ever felt like you're drowning in data? As a researcher studying movement, you probably have. Accelerometers, step counters, GPS, heart rate monitors... the list of sensors goes on, and each spits out data in its unique language. Even the same type of sensor can have different formats depending on the brand. Combining or comparing data becomes a frustrating puzzle, like trying to fit square pegs into round holes.

But wait! Before you resign yourself to a life of data-wrangling drudgery, there's a solution. Think of it as a translator for your data woes: readers (i.e. parsers). These scripts are like superheroes swooping in to rescue you from the mess. They can load data from different sensors and vendors while transforming it into a standardised format. While it may not be perfect, having everything in a consistent structure is a game-changer.

P.S. Welcome to the data-wrangling Hall of Fame. Mapping X, Y, and Z axes from different accelerometer devices worn on various body parts. It's like trying to decipher a secret code.

From Basic Data to Rich Features

We can go beyond just wrangling your data. Do you have GPS data but lack information on distance and speed? Easy! We can calculate that for you. And here's a real mind-blower: with just latitude and longitude, we can determine elevation! No, not by pulling data from some database, but through the magic of satellite imagery and some seriously complex math (don't worry, you're not alone if you don't understand it!).

So, ditch the data struggles and focus on the fun part – uncovering the insights hidden within your movement data!

Taxonomy: Building a Unified Language for Movement Behavior

Standardised data formats are a researcher's dream come true. Algorithms hum along, tools work flawlessly, and everyone's on the same page. But what about a universal language for describing the data itself? A world where "sedentary behaviour" has a clear definition, and "sporadic walking" isn't open to interpretation (unlike that colleague who insists jogging and slow running are… interchangeable? Or are they?).

The confusion is real. Some researchers operate with razor-sharp distinctions between terms, while others navigate a sea of ambiguity. Here's the key point: a slight inconsistency might not be a research deal-breaker, but clear definitions are the cornerstone of building a unified understanding. After all, how can we answer our research questions if we're speaking a different language?

Turning Numbers into Narratives

Sure, a movement behaviour taxonomy sounds fancy – a whole system for classifying those 24 hours of movement data. But for us researchers in the trenches, it's not enough. Understanding the theory behind movement patterns is great, but you're reading this because you need to wrestle with real data, right? The big question is: how do we take all those time series and numbers and turn them into something meaningful? We can't just dream up categories – our movement behaviour taxonomy needs to be a bridge, a way to translate raw data into a clear picture of how a specific person moves and behaves. That's the goal: not just theory but a practical tool to explain behaviour from the data.

Processing: The Missing Piece in Your Movement Behavior Puzzle

The data is ready, the naming conventions are clear, and the link to movement behaviour is established. Ready to crack your research question? Not quite! Just because you have a person's acceleration in X, Y, and Z values doesn't mean you can instantly decipher someone's activity. That's where powerful classification algorithms come into play. These tools help us label our data (remember, in that standardised format we discussed?) and connect it to specific behaviour categories within our taxonomy.

We all know endless debates about which approach reigns supreme won't get us anywhere. Real progress, the kind that gets everyone moving more, comes from teamwork and constantly refining our methods and algorithms. Let's ditch the idea of a single "best" method and focus on making these all truly effective together!

Join the Movement

Do you think you've got a killer algorithm for physical activity intensity? We can help you LABDA-fy it! Plug it into our Python package – it's like the Matrix for algorithms and methods. Now, everyone can join the movement.

Here's the reality: someone might develop an even sharper method someday. But hey, until then, you'll be a champion for healthy habits! You'll be helping researchers understand movement, which translates to getting people moving, building fantastic playgrounds, and getting teenagers off the couch!

From Legacy to Cutting-Edge

LABDA's vision is to evolve into a powerful toolkit overflowing with various algorithms. This expansive approach empowers researchers to delve into a broader range of research questions, unlocking a deeper understanding of human movement. While cutting-edge algorithms pave the way for innovation, we recognise the enduring value of established methods like the 'counts' metric. By seamlessly integrating legacy and cutting-edge approaches, LABDA bridges the gap between past and present research, fostering comparisons that unlock even richer insights.

Analysis: Unlocking Insights

We've covered a lot of ground! Your data is processed and annotated. But is that all you need? It depends on your expertise. Experienced programmers can likely navigate the next steps independently. However, even if you're new to this, we've got you covered!

The core of analysis is generating clear and concise results – summaries you can use directly in your research paper. Visualisations like plots and charts are also included, along with the possibility of creating beautiful reports for your participants.

With this comprehensive toolkit, researchers and individuals can gain valuable insights into human movement.