Ye Olde Future Temper-O-Meter

FACT: Data is awesome, but its awesomeness is usually confined to a digital canvas. We wanted to try showcasing it in a completely new way by using a physical installation to surface sets of time-series data. Our own data and analytics team was keen to chart all sorts of metrics, but we eventually settled on a giant, mechanical, one-of-a-kind installation that moves, illuminates, and constantly updates to show the predicted weather forecast for the next 15 hours at our different office locations.

  • Challenge

    Liberate data from its digital confines and go analog with its presentation.

  • Solution

    A reconfigurable installation that can display everything from weather forecasts and historic GA data to traffic conditions heading west on 70.

  • Results

    A meticulously well-crafted mechanical weather chart that features moving cogs, hands, and six chart segments adorned with old-tyme elegance.

How we did it.

  • Sketch/UX

    We started sketching rough concepts that outlined necessary functionality and UI elements that would help tell the story of time-series data.

  • CAD most things. Measure twice, cut once.

    CAD helps us design custom components and plan a layout so assembly is a breeze. Including printed templates for helping with assembly.

  • 3D printing for production

    We printed everything from custom washers to entire GT2 timing belt tensioning idler pulleys in about a week — a total of 20 unique 3d printed parts. By using a co-polyester like nGen we were able to minimize the amount of post-processing to some filing.

  • Mobile App

    As a final touch we provided users a way to enter a location.

  • Light spool mechanism

    We needed a way to button all the loose ends up -- even the difficult, and literal, loose ends. The LED rope we used was just hanging, so we followed the advice of UX and built a simple coiling mechanism.

  • Charts aren't pretty. It’s hard to convey all the necessary information like axis scale or current position, let alone supplement information such as annotations without it looking like...well... a page of charts. By limiting the type of data that is charted, and controlling the motion of the lights, we were able to convey useful information in an elegant package.

    Inventing the Temper-O-Meter, an IoT Marvel
    Justin Sinichko, Hardware Developer
  • There are two types of movement: physical up/down of the chart’s segments, and the color of the leds. We needed to tackle the problem of knowing which leds to illuminate, and when. Because the chart is physically moving, as well as taking in and giving out LED rope slack, we needed to be smart about it. Math saved the day.

  • The most obviously compelling part of the installation is the movement of 6 segments —  we’re only human, and movement is fascinating to us. So when UX said we needed to tidy up the LED tail with a spool, we went for an approach that utilized a hanging weight that moved up and down to apply a slight and constant pull.