Featured image: Stars Over Mountains - Photo by Casey Horner on Unsplash

The Concept

Friday Night Experiments are those experiments and off-hand projects which actually involve doing something but that are tangential to my “real” work; or experiments I just expect to fail badly.

The term comes from Andre Geim and Kostya Novoselov who frequently held what they termed “Friday night experiments” – sessions where they would try out experimental science that wasn’t necessarily linked to their day jobs.

The Origin Story

Indeed their Nobel Prize-winning work on graphene started out as a Friday night experiment – the phrase used in the Geim lab to describe investigations that were considered slightly crazy. The reference is to the time of week Geim first tried the experiment that led to his 2000 Ig Nobel Prize (levitating a frog).

One Friday, the two scientists removed some flakes from a lump of bulk graphite with sticky tape. They noticed some flakes were thinner than others. By separating the graphite fragments repeatedly, they managed to create flakes that were just one atom thick. Their experiment had led to graphene being isolated for the very first time.

The Philosophy

This playful approach is fundamental to how both Andre and Kostya work. It is seen as both:

• A useful way of maintaining interest
• A means of generating new ideas
• Permission to fail spectacularly
• Liberation from “productivity” metrics

Why This Matters

Academic research often becomes:

• Overly focused on fundable questions
• Risk-averse (publications over exploration)
• Bound by hypothesis and methodology
• Disconnected from curiosity

Friday night experiments reclaim the joy of investigation without predetermined outcomes.

The Approach

My version of this practice involves:

Permission to Be Tangential

• No requirement to link to main research
• No pressure to publish
• No grant applications or justifications
• Just curiosity and possibility

Expectation of Failure

Most experiments fail. That’s fine. Actually, that’s the point:

• Learning what doesn’t work
• Discovering unexpected problems
• Building technical skills
• Generating better questions

Low Stakes, High Play

• Weekend or evening projects
• Modest budgets (personal funding)
• Collaboration with amateurs and enthusiasts
• Documentation without publication pressure

Accidental Usefulness

Sometimes these experiments:

• Reveal interesting phenomena
• Develop into real research
• Build unexpected skills
• Connect to other projects

But that’s a bonus, not the goal.

Current Experiments

BeeLife: Urban Beekeeping and Monitoring

Maintaining honeybee colonies in urban Manchester while monitoring:

• Hive conditions (temperature, humidity)
• Bee behaviour and health
• Honey production and quality
• Environmental factors

Status: Active since 2018
Unexpected finding: Bees as pollution sensors

Honeybees as Pollution Monitors

Using bees as bio-indicators of environmental quality:

• Collecting pollen and nectar from wide areas
• Analysing honey for heavy metals and pollutants
• Mapping pollution patterns across Manchester
• Citizen science collaboration

Status: Pilot phase
Challenge: Analytical chemistry is expensive

IoT Hive Monitoring

Building custom sensors for:

• Internal hive temperature and humidity
• Weight (honey accumulation)
• Sound (swarming detection)
• Image recognition (varroa mite monitoring)

Status: Prototyping
Learning: Bees destroy electronics enthusiastically

The Geim Example

It’s worth noting Andre Geim’s trajectory:

2000: Ig Nobel Prize for levitating a frog (Friday night experiment)
2010: Nobel Prize in Physics for graphene (started as Friday night experiment)

The playfulness didn’t diminish the rigour. The “crazy” ideas led to fundamental discoveries.

His approach demonstrates:

• Serious science can embrace playfulness
• Tangential exploration breeds innovation
• Permission to fail enables risk-taking
• Curiosity matters more than credentials

What Counts as a Friday Night Experiment?

For me, these are projects that:

Are Disconnected from Main Work

• Not about web accessibility (my primary field)
• Not about human-computer interaction
• Not seeking publications
• Not on work time (hence “Friday night”)

Involve Actually Doing Things

• Not just reading or thinking
• Hands-on experimentation
• Building or observing
• Getting messy (literally, with bees)

Have Low Expectations

• Might fail completely
• Might reveal nothing
• Might be interesting only to me
• Might occasionally surprise

Generate Learning

• New skills (electronics, chemistry, beekeeping)
• New collaborations (citizen scientists, local beekeepers)
• New questions (how do cities affect bees?)
• New appreciation (for complexity of “simple” things)

Why Bees?

The bee experiments emerged from:

1. Curiosity - How do hives work? What’s actually happening in there?
2. Tangentiality - Nothing to do with my research field
3. Accessibility - Can keep bees in a garden
4. Depth - Simple question leads to endless complexity
5. Connection - Links to ecology, chemistry, data science, IoT

Bees are perfect Friday night experiment material:

• Endlessly interesting
• Require actual work (not just computation)
• Connect to broader questions (environment, sustainability)
• Have high failure rate (colonies die, experiments fail)
• Generate surprising insights

Lessons Learned

Technical Skills

• Soldering and electronics
• Sensor calibration
• Data logging and analysis
• Beekeeping basics (still learning)
• Failure diagnosis (essential)

Scientific Mindset

• Observation before intervention
• Documentation matters
• Controls are hard in real world
• Nature doesn’t read papers
• Complexity exceeds models

Personal Benefits

• Mental break from usual work
• Physical activity (hive management)
• Different kind of problem-solving
• Connection to natural processes
• Humility (bees don’t care about your PhD)

The Permission Structure

Calling something a “Friday night experiment” creates useful permission:

For Myself

• To fail without career consequences
• To be amateurish (I am an amateur beekeeper)
• To learn slowly
• To prioritise fun and curiosity

For Collaborators

• To contribute without credentials
• To suggest “crazy” ideas
• To question my assumptions
• To expect rigour without publication

For Evaluation

• Success ≠ publication or funding
• Success = learning something interesting
• Success = enjoying the process
• Success = generating new questions

Challenges

Time

Academic life is already overcommitted. Friday night experiments compete with:

• Real research
• Teaching responsibilities
• Family time
• Sleep

Balance is tricky.

Resources

Self-funded means:

• Limited equipment
• Can’t do expensive analysis
• Reliance on DIY solutions
• Scrappier approaches

But constraints breed creativity.

Legitimacy

These aren’t “real” research. That’s both liberating and frustrating:

• Can’t claim work time
• Won’t advance career
• Not recognised in assessments
• But also: no pressure, no justification needed

Scope Creep

Experiments want to expand:

• More sensors
• Better analysis
• Proper controls
• Publishable outcomes

Resisting this is important. The point is playfulness, not productivity.

Future Experiments

Ideas in various stages of “might try”:

Completed Failures

• Automated varroa mite detection (bees destroyed camera)
• Sound-based swarming prediction (too noisy, too complex)
• Temperature-based honey production forecasting (too many variables)

Active Investigations

• Pollution mapping via honey analysis
• Seasonal patterns in hive weight
• Correlation between weather and foraging

Possible Future

• Comparing urban vs. rural hive health
• Citizen science network of hive monitors
• Integration with local school education
• Open hardware designs for beekeepers

The Broader Point

Friday night experiments aren’t just about bees or graphene or levitating frogs. They’re about:

Maintaining curiosity in a field that can become narrow and professionalized

Permission to play when work becomes serious and measured

Different ways of thinking that emerge from different domains

Joy in investigation separate from outcomes and metrics

Connection to physical world for those of us in computational fields

Academic Culture

Universities often say they value:

• Innovation
• Risk-taking
• Interdisciplinarity
• Curiosity-driven research

But reward systems incentivise:

• Safe, fundable research
• Publications in established venues
• Work within defined fields
• Demonstrable “impact”

Friday night experiments are the gap between rhetoric and reality. They’re what academics do when they’re allowed to ignore the incentive structures and just investigate.

The Geim Legacy

Andre Geim won both an Ig Nobel Prize (for seemingly silly work) and a Nobel Prize (for groundbreaking discovery). The same playful mindset underlay both.

This isn’t coincidence. The willingness to investigate “silly” questions, to fail publicly, to work outside your field – these habits of mind enable discoveries that more cautious approaches miss.

Friday night experiments honour this legacy. Not by trying to win prizes (that would defeat the point), but by embracing the same spirit: curiosity, playfulness, and permission to fail in interesting ways.

Conclusion

Most Friday night experiments fail. That’s fine. Some are moderately interesting. That’s nice. Very occasionally, one reveals something genuinely surprising.

But the real value isn’t the findings. It’s:

• The practice of curiosity
• The skills developed
• The mental space created
• The reminder that science can be playful

And every so often, like Geim and Novoselov separating graphite with tape, something unexpected emerges from the tangential, the playful, the seemingly pointless.

That’s worth a Friday night.