The Web Demographic Theory has been put together by Talking Birds' team of data miners. It is a hothaus commission from Vivid, Birmingham's Media Lab, and is supported by Arts Council England, West Midlands.

The theory was developed responsively whilst in residence at Vivid (in the Big Peg in Birmingham UK's Jewellery Quarter) using the words, pictures and sounds submitted by the Web Demographic community of volunteer participant collaborators.

The Web Demographic Data Miners

Data miner JV has spent more person-hours collecting data than either of the other two. As a result, she has become dangerously attached to many of the participants with whom she has been conversing, and she is particularly keen to ensure that the theory is nicely presented, so that her new friends won't be disappointed.

Data miner NW has been lauded for devising the already legendary 'Random Raffle' (Pat. Pending) method of data analysis. The results of this new method of data analysis are potentially far reaching and have been particularly useful in reducing the number of data nodes and distilling the facts throughout Web Demographic.

Data miner DN has been the calm voice of reason throughout the data collection and analysis. His keen ears and preoccupation with sound have been invaluable in identifying and classifying data relating to the (tangential) ice cream war in the Jewellery Quarter. He is also responsible for naming and explaining the 'Kinneval Anomaly'.

The Web Demographic team of data miners works under the collective name of Talking Birds. If you would like to contact the team, you can email us by clicking here.

View diary of Talking Birds' residency at Vivid


Many of the images on this site have been submitted by and are used with kind permission of Web Demographic participants.

The Web Demographic Participant Collaborators

Throughout Web Demographic, Talking Birds has been engaged in a number of email conversations with a burgeoning community of participant collaborators. These conversations have, on the whole, been protracted, un-prescriptive and extremely interesting.

Participant emails have determined the direction in which the Web Demographic theory has developed, and the eventual content of the ten statements of the theory are entirely derived from Talking Birds' analysis and filtration of the
words, pictures and sounds submitted by members of the Web Demographic community.

Our volunteer participant collaborators are listed below (in alphabetical order);

Lizzy A
Rachel A
Julian B
Jennifer C
Sue C
Andy D
John D
Kay D
Cormac F
Tom F
Isla F
Jon F
Julie F
Rose F
Caroline F
Alex F
Arnim F
Chris G
Golden G
Clive H
Johnny H
Isabelle H
Rachel H
Saul H
Sarah HS
Mark HS
Eleanor H
Clare H
Jin G
Kate FS
Alex K
Debbie K
Nadia K
Mike K
Maurice M
Steve M
Tom Mc
Mick M
Lisa M
Mervyn M
Floyd M
Simon M
Carolyn M
Amy N
Dave N
Nicola (A)
Jim N
Vanessa O
David O
Nilam P
David P
Roxane P
Matt P
Nicola P
Pixie Q
Jeremy R
Nicola R
Nick R
Alan R
Graeme R
Betty R
Anda R
Benny S
Jack Sh
Jack Si
L Stevens
Craig S
Helen S
Paul S
Jennie T
Dan T
Nick T
Jimmy T
Lucy W
Alan VW
Vicky W
Kaye W

Many thanks to you all.

logo of arts council england, who funded the project

Back to Credits | About Data Miners | About Participants

Residency Diary

Go to:
Week 1 April 4th | 7th | 8th | 9th | 11th
Week 2 April 14th | 15th | 16th | 18th
Week 3 April 21st
Week 4 April 28th
Week 5 May 6th

04.04.03 Week 1

Three data miners take the A45 to Birmingham (45% traffic lights were green). Incompatible laptop computers will cause anomalies in the parametric regression procedure. Awkward silence in the car.

87% car park spaces were full. Ceilings low. Web Demographic vehicle breaks its aerial. 93% radio stations lost. Including the one carrying car park height information.We park on the roof. Weather: windy. Data lost: 2 pieces of low-level categorical dependent variables (non essential).

Key card no longer works in the lab door. Seems it is programmed remotely by a company in Hong Kong whose computers have been sprayed with strong disinfectant and have crashed causing corrupted codes. All over the world people cannot access their premises. JV bypasses the system using a magnet and piece of packing tape.

View from lab window: GM International, Jewellery Distributors. Hockley Trade Centre, Jewellery Wholesalers. Alan Stevens, Ornamental Jeweller. Carol’s Discount Jewellery Warehouse. There are enough rings produced here for three on every finger of every Birmingham resident.

The Web Demographic data is complex. Which binary recursive partitioning rule to use? JV favours Gini, DN argues for Entropy, NW pushes hard for Twoing. Sample data is used with all three systems. Twoing creates decision trees of such beauty that NW actually cries. Emotion cannot influence findings and so Gini is elected by a count of two votes to one. NW goes missing for two hours.

Today’s finding: an increase in spare time is seen as a threat to the mental health of 12% of web demographic participants (complexity parameter 0.00112).

back to residency dates | back to top of page


NW does not report for work until Midday; 8.655% of his pay is docked. He gives an excuse which is believed by 0% of the Web Demographic team. ‘But I’ve got kids to feed,’ he wails. He is advised to take the matter up with Off-Web, the Web Demographic regulator. NW spends all afternoon composing a letter. More pay is docked. He drops the matter before he loses his bus fare home.

Attachments are printed out and put on the walls. The people in them become the data analyst’s family. (Over) Excited by the response from the Web Demographic community to phase 2, JV and NW have already started to talk about phase 3. DN is still nostalgic for phase 1 and is finding it hard to move on. He mutters his sorrow to a picture of someone’s cocker spaniel.

Weather outside: blue. Some smoke is seen in the distance (a fire? an explosion?). Some of the jewellery outlets have closed down. Citizens don’t wish to flaunt their decadence in these troubled times. A plane flies over head. Is it carrying missiles? DN takes a digital photo but it’s hard to see if there really are missiles or just rogue pixels.

We analyse fast using a random raffle ticket model. The Power Point presentation ‘The Hybrid CART-Logit Model in Classification and Data Mining’ was awarded “Best Presentation” at the Eight Annual Advanced Research Techniques Forum, American Marketing Association, Keystone, Colorado, 1998. We here at Web Demographic vow to win this award in 2003. The day finishes with primal whoops and a flurry of raffle tickets.

back to residency dates | back to top of page


Data Miner NW rings in sick at 4am suffering from insomnia. Data Miner JV suggests he starts mentally pruning the branches off his maximal tree one by one, which always works for her. After that he rings in with half hourly updates.

The entire team is now complaining of sleep deprivation but, on arriving at the Lab, is fuelled up by a healthy number of ‘Phase 2’ responses that have come in overnight. Though a mixture of random raffle selection, free-form associative thinking, and Worthing Zen, a draft of the first theorem begins to emerge:

Finding pleasure in the mundane is an essential part of modern life. [Click here to let us know your opinion on this]

Another common thread that seems to be emerging is Canada. [Click here to let us know if you too have a link to Canada]

"We are all Web Demographers now!" shouts Data Miner DN from the roof of the Big Peg in a moment of elation, fuelled no doubt by the 25 cups of coffee a day recommended by one respondent.

back to residency dates | back to top of page


Data Miner JV is on a solo shift at the lab today, and finds herself slightly overwhelmed by the sheer numbers of emails responding to Phase 2 questions, which are still coming in thick and fast. She calculates that it takes an average of 4.2 minutes to file and deal with each message. The Lab is now so full of paper that she makes a mental note to enquire about taking over the edit suite next door as a data storage annexe.

There are also emails coming in confirming further links with Canada, and JV discovers that two more common threads are rows of doors, which are the same shape but different colours [click here if you have any experience of rows of differently coloured doors.]; and the perception that space equals productivity. [Click here if you have an opinion to give us on this.]

Braving a strange smell in the lift, Data Miner JV dips out to buy a cheese and tomato melt baguette from the exceedingly friendly shop downstairs (23% of the melted filling has escaped from between the bread by the time the sandwich is eaten) and then gets set into a geographical survey of the spread of respondents which promises to take most of the rest of the day.

back to residency dates | back to top of page


Three landmark buildings which were visible from the window of the Web Demographic lab have now disappeared. Two theories emerge: one, we have been here for much, much longer than our diary dates suggest; two, we have come to the wrong city today. In order to resolve the issue, NW runs a parametric model using classification and regression trees. The computer churns out the following conclusion: results are invariant with respect to monotone transformations of the independent variables. NW asks for clarification. The computer says: pull yourself together.

A picture of a pageant queen causes an anomalous node on the classification tree. If the picture is of Miss Colorado we have a high probability of achieving a closed and satisfying hypothesis. If it’s Miss West Virginia then we’re pretty much screwed. We ask the Web Demographic participant who sent the attachment for clarification but our email to him is, apparently, undeliverable. NW suggests writing ‘Miss Colorado’ on her red sash using digital manipulation techniques. This produces shocked intakes of breath from DN and JV. NW tries to make a joke out of it but soon realises it’s no laughing matter. Faking results could get the lot of us chucked out of the building. He keeps his mouth shut for the rest of the day and isn’t offered a lift home.

A familiar ice cream jingle is heard from the Web Demographic lab. Mr Gale? Has Mr Gale strayed 23 miles from his South Coventry patch? We look out the window. It’s not Mr Gale. It’s Mr Super Soft using the same jingle. We try to tell him that there are over 350 ice cream van chimes available to the mobile ice cream seller and there’s no need to duplicate and thereby antagonise Mr Gale. Mr Super Soft tells us to keep our gobs shut, and plays his jingle for more than the legal time. DN attempts to impound the van and perform a citizen’s arrest. He is dragged at slow speed down the street.

back to residency dates | back to top of page

14.04.03 Week 2

The theory is beginning to take shape, and our team of Data Miners have been sending draft statements out to participants for comment - 95% of participant-comments received so far appear to back up the team's findings, which is a relief.

There is a slight nervousness in the air about opening the theory up for scrutiny as some of the findings are quite unexpected, but as data miner DN points out - it's all based on the findings of the research, so we can back it up, even if it seems slightly implausible.

Reassured, the team gets back to work, trying not to worry about the niggling "Kinneval Anomaly".

back to residency dates | back to top of page


The majority of the initial data analysis complete, the team is now starting to concentrate on devising methods of demonstrating the theory to the wider public. Data miner JV is charged with translating all the information into a form that will be viewable via the internet.

This is proving to be a slightly more challenging task than originally envisaged. She only manages half a day before falling to the floor in a dramatic faint.

back to residency dates | back to top of page


Data miner JV is working alone again, preparing for working in Coventry from Friday since Vivid is closed all next week. In order to get into practice she works in Coventry for the first half of the day.

Later, in a Birmingham suburb, JV hears an ice cream salesman using the theme tune to "Match of the Day" as his jingle. She nearly jumps a red light whilst trying to locate the offending van.

This somehow prompts her to think that turquoise might be a good base colour. [Perhaps something like 99FFFF]

back to residency dates | back to top of page


The team of three reconvenes in Coventry to look at the website proofs produced by Data Miner JV, and discuss further the format of the theory's web demonstrations.

Participants are still returning their opinions on the formulation of the theory. Several interesting opinions have been received on the matter of Point 6 about the authenticity of families in old photographs. Participant Dan reckons that it's the old families that are more authentic and that the age of the photograph has no bearing on the matter. Tom, meanwhile agrees resolutely with the theory, whilst Jack posits that families where there are no photographs at all may in fact be the most authentic.

The team muses on these emails, but for the time being stands by its original theory.

back to residency dates | back to top of page

21.04.03 Week 3

Data Miners NW and DN are fed up with the fact that Data Miner JV hasn't finished the website yet. They decide to lock her away to make sure she gets on with it, and refuse to let her out until next Friday afternoon.

Since they are both boys, neither of them can work a computer and so Talking Birds regrets that there will be no more diary entries for the remainder of this week.

To complain about the discontinuity in the diary service or about the mistreatment of Data Miner JV, click here.

back to residency dates | back to top of page

28.04.03 Week 4

The team of three Data Miners have reconciled their differences during a furiously-paced week of data processing. (Data Miner JV would just like to thank all those of you who emailed to protest at her confinement in the cupboard.)

We are now entering the final third of the residency at Vivid. Mindful that Vivid are due to kick us out of the residency lab in a couple of weeks time, our team of Data Miners will be keeping a low profile for the next 5 days, making the most of the time left. Our best estimates suggest that we are on track, having completed approximately 60% of the presentation work so far.

Data Miner JV will be at Vivid on Monday 28th and Wednesday 30th, the whole team will be there on the morning of Friday 2nd. Do drop in and make us a cup of tea if you are passing The Big Peg...

6.5.03 Week 5

Well, this is it, the time has come to unveil the theory to an unsuspecting public. Nerves are frayed, but spirits are buoyant. The team is 100% behind the theory and in bullish mood as they prepare to publish the site. The last minute tweaks are all that remain to be done. Time to close the current research site and put up a holding page to inform people that the publish date is this Friday - 9th May 2003 - a day that may go down in history...

Back to top | Back to Residency Dates
examine the theory process, method and rationale data on participants send us your feedback on any part of the theory view site map research credits frequently asked questions
web demographic [title bar]
examine the theorylink to frequently asked questionslink to research creditslink to site maplink to details of participantslink to process, method and rationalelink to web demographic theory