TINK - sketching product experiences of connected objects

86 - TINK
5.2 Connecting a coffee machine to the
Internet
The first test was aimed at understanding
how to connect an Arduino YUN to the
Internet via Web-Socket.
The first trial system was installed on the
coffee machine at the StudioLab. The idea
was to hack the coffee machine to measure
it’s usage during the day and visualize the
data on a web-page.
Developing this prototype was also a
good way to test how much a simple data
visualization can help discovering patterns
and information that otherwise would not
be evident.
A web NodeJs server based on Express
was developed. The server implements
Socket.IO in order to receive the data from
Arduino. The data is stored in a MongoDb
database and visualized on a public webpage
using a well-known JavaScript data
visualization library called d3js.
All the project’s code and the Arduino
library’s installing instructions are
published on GitHub and are retrievable
at https://github.com/lorenzoromagnoli/
SocketIOYunClient.
A proximity sensor was hidden inside the
coffee machine(Figure 40). The sensor was
able to detect if a cup was placed in front
of it and therefore count how much the
coffee machine was used along the day.
The sensor was connected to a battery
powered wireless Arduino board, developed
inside the StudioLab, called Schizzo (Figure
40). The Schizzo was connected via Xbee
to an Arduino YÙN (41) connected to the
Internet network of the faculty via WIFI.
An Arduino library was developed to
connect the board to a custom web server.
The developed library, modeled on the
SpaceBrew library for Arduino YUN, makes
use of Socket.IO library to provide real-time
bidirectional event-based communication
to the Arduino YUN. Socket.IO was chosen
as it works on every platform, browser and
device, focusing equally on reliability and
speed.
This trial ran for a couple of days recording
live data of the StudioLab’s consumption
of coffee. The data was publicly visible on
a website for the length of the experiment.
The screen-shot presented in Figure 42 is
a representation of the experiment, the
bar-chart grouped the amount of coffee
that was made be the machine every 15
minutes.
From the visualization it was possible to
identify trends on the usage of the object:
People arrive to work in the morning
between 8.30h and 10h; the consumption
of coffee keeps on growing until lunch,
when it drops because of the lunch break;
after lunch, the next peak is reached around
16h; after that, the amount of coffee keeps
on dropping until the time people leave
the office, around 18h.
What is also interesting to observe from
the data visualization is how the coffee