@datagrok/pedometer NPM

Pedometer is a package for the Datagrok platform. It showcases the ability to develop plugins that include scientific computations and a rich UI that leverages the platform's capabilities.

Here are the files of particular interest:

scripts/detect_steps.py : a Python function that detects steps based on raw accelerometer data.
package.js : a JavaScript function that accepts columns containing x, y, and z acceleration components, calculates position of steps by invoking detect_steps, visualizes the raw data on a chart, and customizes chart's interactivity.

Detecting steps

Let's take a detailed look at the header of the scripts/detect_steps.py file. It addition to containing information such as name and description, it instructs the platform what the input and output parameters are, including their data types and semantic types:

# name: detect steps
# description: detects positions of steps based on accelerometer data
# language: python
# tags: template, demo, accelerometer
# sample: sensors/accelerometer.csv
# input: dataframe accel [accelerometry data table]
# input: column x {semtype: accelerometer-x} [x axis]
# input: column y {semtype: accelerometer-y} [y axis]
# input: column z {semtype: accelerometer-z} [z axis]
# input: double sample_rate = 32 [sample rate, in hz]
# output: dataframe steps {action:join(accel)} [steps positions]

Javascript

//input: dataframe accel [Accelerometry data table]
//input: column x {semType: Accelerometer-X} [X axis]
//input: column y {semType: Accelerometer-Y} [Y axis]
//input: column z {semType: Accelerometer-Z} [Z axis]
//input: column timeOffset {semType: Time-Offset} [Time offset column]
pedometer(accel, x, y, z, timeOffset)