“When working on data-driven engineering applications I’ve often felt I was missing an easy and convenient way to visualize and interactively select data”, says Team Lead for Controls & Embedded at 1X, Manuel Yves Galliker.
Consequently, Manuel decided to develop a bespoke solution to address this need, which he named The Visual Pandas Selector. Beyond his role as a robotics engineer and tinkerer, Manuel is also involved in a range of open source software projects.
The Visual Pandas Selector is a tool to visually select portions of numeric time-series data from a pandas dataframe. It’s intended to provide a fast interactive way for manual data selection, which can be very useful in many applications including machine learning, statistics or system identification.
“It's rewarding to see something I've made being useful to others. Especially since much of my software work owes its success to the contributions that countless developers have generously made to the community”, says Manuel, adding that the python package is publicly available on GitHub and can easily be integrated into an existing pandas DataFrame workflow.
The user can effortlessly configure the tool to plot dataframe columns in vertically stacked subplots and view data distributions with the included histogram feature. With a simple click-and-drag action, horizontal data windows can be selected, and the tool will automatically combine them into a new dataframe.