The Prediction Machine
Rachel Jacobs (Artist)
Since October 2014 Rachel Jacobs has presented The Prediction Machine, with over 800 people interacting directly with the machine during exhibitions alongside a series of public workshops that have taken place at Nottingham Contemporary, Loughborough Town Hall, Loughborough University, the University of Nottingham.
The Prediction Machine is an interactive artwork, based on end of the pier fortune telling machines. The machine marks ‘moments of climate change’ in our everyday lives and prints out ‘climate fortunes’ for 30 years in the future, that visitors to the machine can take away with them. These predictions use live weather data captured at a local weather station, matched with projected climate data from future climate models provided by scientists at the UK MET office, and observations by local people. The machine links up to an interactive website that combines narrative and visual representations of the data with more traditional science communication.
The Prediction Machine has been developed in collaboration with local people in the East Midlands (UK) and Rio State (Brazil), engineers, computer scientists, climate scientists and researchers.
The Prediction Machine’ explores how performances of data can occur through the interaction with a machine that tracks climate data and marks moments of climate change.
An early study of the The Prediction Machine reveals that participants in the public workshops that took place alongside the exhibition of the machine were invited to compare personal, sensory observations with scientific data, by observing the weather and enacting scientific processes for sensing and modelling the data. They used this experience to write predictions for that day 30 years in the future – taking into account scientific climate models of projected temperature increases. These participants appeared to be engaged in a process of ‘performing’ and ‘interpreting’ the data as a physical, logical, emotional and sensory experience’.
Users of the machine in the exhibition were then invited to take these predictions away with them as souvenirs of the moment in time when they interacted with the machine, requiring them to reflect for a moment on the weather outside and imagine a future climate scenario based on that moment. The exhibition was designed as a journey of discovery – from powering the machine, taking the print out, going to the Promises and Wishes machine to input your own ‘promises, wishes and predictions’ and then finding out how the machine uses the data. This journey turns the visitor into part of the performance of data on the machine and the process of modelling future climate change.
The Prediction Machine is an interactive, playful and performative artwork designed to represent live weather and future climate data. The data is presented on a screen, visualized through a lit up sign and printed out as narrative based predictions informed by the scientific data. Temperature, precipitation and wind speed data is recorded at a local weather station every 5 minutes and fed to the machine via the Performing Data WordPress Plugin – Timestreams.
Python code on a Raspberry PI sits between the weather station and the wordpress plugin. This records the weather data and transforms the live data into (a) projected temperature data for 30 years in the future, based on a simple model proposed by climate scientists (b) narrative weather scenarios written by the artists, that represent the combined live (c) narrative descriptions of the weather as either ‘extreme’, ‘unexpected’ or ‘expected’. This code uses the Timestreams API to send the data to the WordPress plug-in. Here the artist is able to scale the data, replay archived data if a problem with the weather station arises, juxtapose aggregated data (in this case the average monthly temperatures from that location from the last 100 years) and play several ‘streams’ out at the same time, each with the same timestamp so that they appear on the machine and on the project website simultaneously. These streams are then pulled, using the Timestreams API, into a Unity interface running on an internet connected laptop hidden within the machine that controls the lights, visuals on the screen and the printer commands for an industrial kiosk printer also embedded in the machine.
The machine print outs the narrative predictions in response to the live data that appears in a slot on the bottom half of the machine. Unity accesses a .csv file containing all of the narrative predictions allocated to each of the weather scenarios, a random choice of this set of narratives is then printed on a card, alongside a unique web code (the Unix timestamp from Timestreams). The lights are controlled by an ardiuino that gets temperature data from Unity, when an ‘extreme’ reading comes from Timestreams then the lights pulsate as a warning sign. Unity also gets data from a hand crank that generates voltage to power the screen and speakers that play the audio track from the video messages that appear on the screen. This is also controlled through the Arduino.
Additional to The Prediction Machine is a companion piece, the ‘Promises and Wishes Machine’. Visitors can add their own ‘promises, wishes and predictions’ that are uploaded to Timestreams as message text and timestamped alongside the live temperature, precipitation and wind speed data coming from the weather station. This data is available to support qualitative research on how visitors are using the machine. These web pages include a form for users to submit these messages, choose their date of birth from a drop down menu and type in the unique web code they received on their prediction print out. Submitting this form allows them to view two graphical visualisations that use a combination of aggregated and live data from Timestreams.
Both machines are designed in the same materials and style (sustainable oak and aluminium) but this additional machine simply contains an iPad that displays interactive web pages from the project website: http://www.timestreams.org.uk/promises/promises.html
Artist’s sketch of the technical specifications for The Prediction Machine
This artwork was created by Rachel Jacobs in collaboration with Matt Little, Ian Jones (Sherwood Wood), Matthew Gates, Robin Shackford, Juliet Robson, Dr Candice Howarth, Dr Carlo Buontempo, Mark Selby and researchers at Horizon and the Mixed Reality Lab. Funded by Horizon Digital Economy Research (EPSRC EP/G065802/1) http://www.horizon.ac.uk/ Arts Council of England, EPSRC Impact Acceleration Account, and Radar LU Arts.