Overview

A physical data visualisation gets participants exploring datasets with physical materials. Instead of bar charts and graphs, they might make a model of a hybrid animal out of cardboard, create mini trees out of plasticine etc.

Before

Prepare datasets

Data wrangler prepares 3-4 datasets. Each is printed out, generally in spreadsheet form. A good dataset is:

• small enough that any activity can be easily done by hand
• simple enough that calculators are not needed
• interesting

Possible datasets:

• trees (50 trees in a small patch of Melbourne)
• trees (aggregate stats)
• pet registrations
• https://github.com/FloHu/VizBi-HUB

Print out 4 copies of each. Each group will only do one dataset. They will often work on some subset of the dataset (eg, there might be 200 pubs but they pick out the ones that are on Lygon St)

Dataset spreadsheet.

Get data vis inspiration materials

Shopping for materials

Materials:

• balloons
• blue tack
• cardboard
• pipe cleaners
• markers
• paper
• plasticine
• string
• highlighters
• coloured popsicle sticks
• sticky tape
• straws
• scissors

On the day

Set up tables to seat groups of 4-6. Place materials on a central table so participants have to keep moving to get new materials.

Session plan

Intro

Intro: Host explains the activity, the rules, code of conduct etc etc. A broad audience may not be familiar with data visualisation so thoroughly explain what it is.

Warm up activity

• Get people to split up into different parts of room according to criteria like "everyone who works in government/private sector/education", or "born in Melbourne/Australia/overseas" etc.

• First data vis: get everyone to line up in order of age, length of name, etc (hold up fingers in length?)

• Designate the centre of the room as where we are now, then get everyone to go and stand "where they were born/currently live etc". They have to discuss and agree on scale, orientation etc.

• Give groups tiny datasets like "last 3 football scores" or "last 5 two-party preferred poll results" to visualise. Each on a piece of paper/card. They can't make line charts or bar charts, but could use facial expressions, postures etc.

• At the end of this activity participants will:

• Have moved physically around the room and started talking to each other

• Accepted that data exploration can be quick and messy

• Have got away from the idea the column graphs and pie charts are the only way to represent data

Present the datasets

Participants divide into groups of 4-6.

Host explains each of the candidate datasets.

Each group has 2 minutes to decide which dataset they want to work on. Host hands out the chosen dataset to each team.

Don't allow people to spend too long choosing.

Start playing

Facilitators are on hand to help steer them towards progress, possibly giving them suggestions. Prod them to make sure they're not making bar charts.

• Present work in progress
• Present final result
• Final propaganda

Vague schedule

• 5-10 minutes: waiting to start
• 5-10 minutes: explaining process
• 10 minutes: warm-up
• 10 minutes: presenting data sets
• 3 minutes: teams choosing data hand out data sets and equipment
• 25 minutes: playing with visualisation
• 10 minutes: check-in for status update
  • which parts of the data set are you representing?
  • what materials are you using?
  • how are you representing each data characteristic?
  • what story is your visualisation telling?
• 15 minutes: playing with visualisation
• 5 minutes: wrap up playing
• 15 minutes: presentation