This document summarizes a presentation about how data could transform the arts sector. The presenter discusses 7 trends where data is already transforming other industries and could have significant impact on the arts: 1) Personalization, 2) Outcomes-based predictions, 3) Big data, 4) Combining data sets, 5) New collection methods, 6) Consumer access, and 7) Data as art. For each trend, the presenter provides examples of arts organizations that are using data in innovative ways and the impact it is having. The goal is to inspire the audience to find new opportunities to use data in new ways at their own organizations.
3. Think expansively
about how data
could transform
your organiza?on.
I want YOU to spend the next 90 minutes….
To help you do that, I’m mostly going to give you:
• A high-level overview of how data is being used across arts organizations
• An introduction to arts organizations you can connect with for more info
Where a lot of sessions at the conference so far have gone deep, I’m going to go broad on this topic.
Don’t get caught up in the details. Instead: find one new opportunity to use data in a new way at your org.
4. Agenda.
Making the case for data.
7 trends where data is transforming the arts.
7 elements of a data culture.
Q&A.
• I don’t think I need to make the case to many in this room about the value of data, but I’m going to try to do it anyway.
• I see 7 trends
• Because consultants like things in neat sets, I’m also going to give you 7 elements
• I hope to have time for Q&A
5. Why is data so
important anyway?
• Arts sector → data has long played a role in marketing. Fundraising is on the upswing, and idea of measuring for social impact is gaining traction.
• But there’s more ways arts orgs are using data than you might think.
• Museums are probably furthest ahead in the arts sector when it comes to using data.
• I think the change in theatres isn’t happening quickly enough, or dramatically enough.
• Data could be deeply transformational to theatre business models.
6. It’s not just for
technology companies.
Healthcare.
Educa8on.
Government.
Poli8cs.
Media.
Sports.
Arts.
It’s already been transformation to these sectors.
Some of the ideas for how data will transform the arts are lifted directly from examples in these sectors.
7. Revenue.
Single 8ckets.
Individual dona8ons.
Ins8tu8onal
giving.
Board giving.
Other
contribu8ons.Other earned revenue.
Subscrip8on
8ckets.
This is the average distribution of revenue for a TCG theatre in 2015.
I think the pie needs to get bigger overall.
Data trends could have a significant impact on every one of these revenue streams.
We’re going to come back to this chart & see which trends are having the most impact on each stream.
8. Expenses.
Ar8st payroll.
Produc8on costs.
Opera8ons.
Marke8ng &
development.
Administra8ve payroll.
Produc8on
payroll.
This is the average distribution of expenses for a TCG theatre in 2015.
Data trends could have a significant impact on every one of these expense streams.
Data could increase & decrease expenses.
9. Data could have a
significant impact on
cri?cal aspects of your
mission.
Equity, diversity, and inclusion.
Community engagement.
Social impact.
Aesthe8c value.
Data’s impact is on more than just revenue and expenses though.
I see opportunities for data trends to impact some of the big issues in the sector right now.
10. Let’s get specific.
My argument is data:
• already seeing radical transformation in other nonprofit and for profit sectors
• is used in the arts now to an extent, but we have a real moment of inflection
• has the potential to transform arts revenue, expenses, and mission-critical events
• Convinced?
Let’s get specific.
11. 7 trends where data is
transforming the arts.
1. Personaliza8on.
2. Outcomes-based predic8ons.
3. Big data.
4. Combining data sets.
5. New collec8on methods.
6. Consumer access.
7. Data as art.
We’re going to look at several case studies for each of these trends
Remember to think about how these are, or could, impact YOUR organization.
12. Trend 1:
Personaliza?on.
Personaliza+on (in product, price, or
prac+ce) leads to increased efficiency
and/or be9er outcomes.
Data allows organizations to know more about the audience, and therefore adapt the product, the price, or how they are reached, to best fit each person.
This personalization lets you increase revenue, decrease expenses, or otherwise achieve better outcomes.
On the road to personalization, you need to collect data, store it in a structured way, and then actually use it. we’ll see examples of each of those.
13. Collect the data.
Audience research tools are varied and
numerous.
Trend 1: Personaliza0on.
In order to personalize, first you need to get to know your audience. Most theaters now do at least some audience research some of the time.
A few examples:
• Anne Frank House last summer conducted 16,000 “microsurveys” & wrote great blog post detailing findings. Cheap, quick, and you can do it w/o consultants: https://medium.com/@lottebelice/what-the-
anne-frank-house-learned-from-16-000-micro-survey-responses-10bb9884da19
• At Measure, we work with P5 here in town to use social media advertising to test audience awareness, perceptions, and interests. Starting to build profiles of audience segments.
• Many consultants help arts orgs with audience research methods like structured surveys, ethnography, focus groups, data mining (including Measure, Slover Linett, WolfBrown, etc)
14. The English National Opera used a tool called Hotjar (cheap, w/ nonprofit pricing) to better understand what users needed from their website. After collecting feedback from 5,000 users, they killed their
homepage. Go to the site now and you’ll find you can book tickets directly from the homepage. Ticket sales are up.
https://substrakt-live-assets.s3.amazonaws.com/2017/05/16105446/Substrakt-Digital-Works-2-AB-slides-100517.pdf
15. Store the data.
Customer Rela+onship Management
(CRM) databases store a 360* view of
cons+tuents.
Trend 1: Personaliza0on.
The more data you start to gather about individual constituents, the more it helps to have a place to store it in a structured way. CRMs store that data to better communicate, market, fundraise, engage, and
serve constituents.
DMA Friends was one of the first “visitor tracking” programs in the arts field launched 2013. More than 100k friends now, w/ avg 1,000 data points per person - online & on site. Free membership —> diversity
& scale —> donor support.
rjstein.com/portfolio/dma-friends/
16. PCS Playmaker - just launched first in the country for a performing arts organizations. They are giving a session right now about the initiative.
https://pcsplaymaker.org/Login-Signup
17. Use the data.
Adapt the price, the product/experience,
or the (marke+ng/fundraising) process
based on the data.
Trend 1: Personaliza0on.
So now that you have the data, in a place where you can extract & analyze it, now it’s time to use it to personalize.
TRG Arts gave a session earlier during the conference about their work in patron loyalty & dynamic pricing and the really substantive gains they can make in ticketing & fundraising, based on personalization.
https://www.trgarts.com/Whatwedo.aspx
18. SFMoMA is in the midst of user testing their interactive displays (on the table) to be “Eyes-Free Mode” for use by audience who are low/no sight.
Using voice commands to navigate, voice over to describe, cues for where & how to touch the screen.
Personalization isn’t only about making more money, it’s also about better serving diverse user needs.
mw17.mwconf.org/paper/whats-right-with-this-picture-how-a-graphical-interface-for-modern-art-is-meeting-the-touch-of-non-sighted-visitors/
19. Radical future.
The NeMlix model of disrup+on: A new
compe+tor is created when a +cket
service provider launches a theatre
venue/touring opera+on based on their
aggregate data.
Trend 1: Personaliza0on.
Here’s my version of a radical future when it comes to personalization.
Netflix had the opportunity to personalize not just the experience of the product by showing you suggested moves, but later moved into personalizing content as well, based on viewing habits.
No one trusted Netflix to be an innovative tv/film studio, and yet now won 6 Emmies, 4 Grammy noms, and won an Oscar.
Could Tessitura do the same?
What could your organization do differently if it knew 1,000 points of data about each of your constituents?
20. Trend 2:
Outcomes-based
predic?ons.
Outcomes-based predic+ons leads to
increased efficiency and/or be9er
outcomes.
Our second trend is around:
• identifying a specific outcome you want,
• predicting how to get that outcome,
• testing your prediction, and
• *using what you learn* to achieve better results the next time.
21. Trend 2: Outcomes-based predic0ons.
Predic?on.
Models use data about previous behavior
to predict future behavior.
We’re all pretty familiar with this when it comes to fundraising & dynamic pricing for ticket sales.
• UMS (at University of Michigan) is using machine learning to predict audience preference for performances based on 190,000 ticketing transactions: https://arxiv.org/pdf/1611.05788.pdf?
• KC Rep TCG session on Thursday re: measuring which performances mattered
• Slated is a film investment platform that I consulted for previously that created a scoring system to predict investment value for unreleased films based on team, script, and early audience traction: https://
welcome.slated.com/
• ArtRank & Arthena predicted value of visual art investing
• Obama’s Nudge Unit using behavioral economics to predict responses: https://sbst.gov/
22. Alley Theatre has been using a dynamic scoring model since 2014 for fundraising.
Gives a ranked score for every household for 4 models - donor acquisition, major gift, planned giving, capital campaign
Brought in $200k/year for fundraising; plus $300k subscription renewal in its first year.
Now automated process in Tessitura —> the model learns from its success & mistakes each year to re-score patrons.
http://www.centerpointenergy.com/en-us/Documents/PowerTools/Best%20Practices%20in%20Prospect%20Research%20and%20Analytics.pdf
www.btfri.com/assets/Data_and_the_Search_for_Big_Donors.pdf
23. Trend 2: Outcomes-based predic0ons.
Tes?ng.
A/B tes+ng compares 2 slightly different
varia+ons of the same offer to iden+fy
the strongest approach.
A/B testing is being used on websites, in email campaigns, social media advertising, event print mail pieces. The tools to implement A/B tests are very cheap or free, and easy to learn and use.
V&A Museum are using A/B testing on their website to increase ticket sales, and other conversions (email signups, reading time, etc).
www.vam.ac.uk/blog/digital-media/thinking-small-how-small-changes-can-get-big-results
24. Wolf Trap Center for Performing Arts worked with Capacity Interactive to test these two versions of a mobile optimized ticket purchase page. half audience got #1; half audience got #2.
The test version (on the right) increased ticket sales by 19% during test period.
ideas.capacityinteractive.com/you-should-ab-test-that
25. Trend 2: Outcomes-based predic0ons.
Impact
measurement.
Iden+fying and measuring outcomes
enables a more tailored approach to
inputs.
By naming the impact you’re looking for and collecting data on what leads to that impact, you’re better able to adapt your approach and only spend time & money on those activities proven to actually work.
Session concurrent to this one w/ TheatreSquared is looking at measuring increases in empathy after being exposed to theatre.
Oakland Museum of Contemporary Art is exploring how the museum contributes to an empathetic and caring community, look for blog post soon about their process. https://medium.com/new-faces-new-
spaces
CultureMetrics (UK) is looking at definitions of quality of art and how to measure them: www.culturemetricsresearch.com/
26. The Happy Museum Project in the UK is great to look at how museums are thinking about measuring wellbeing, resilience, civic engagement, and environmental sustainability of: museum go-ers, museum
staff, and communities. This is just one page of a long document about the protocols they’ve set up to measure impact on these outcomes.
happymuseumproject.org/
27. Trend 2: Outcomes-based predic0ons.
Impact investment.
A known, valuable, predictable (social)
outcome can be invested in.
Not only does impact measurement allow organizations to improve their own outcomes, venture capital investors are becoming interested in investing in those organizations who can prove they’re effectively
meeting those outcomes.
Fractured Atlas recently announced their working on an Exponential Creativity Fund to match artist entrepreneurs with venture capital investors to scale new business opportunities. Opens up a significant
new revenue stream for artists.
https://blog.fracturedatlas.org/adventures-in-impact-investing-march-recap-ab61117974da
Philanthropies are shifting their investment dollars to social impact as well (Heron Foundation)
https://www.philanthropy.com/article/Impact-Investing-Pioneers-Take/240182/
28. Social entrepreneurship & impact investing are well established in other fields, but still by & large siloed from the arts.
Upstart Co-Labs opened last year to connect artists with social entrepreneurs, impact investors, social enterprises, and sustainable companies to unlock new capital to scale their creative approaches.
Using data from the social impact created by artists to make the case to new types of investors.
www.upstartco-lab.org/
29. Trend 2: Outcomes.
Radical future.
The poli+cal campaign fundraising model
of disrup+on: Theatres see significantly
increased numbers of small dollar donors
aVer making outcomes-based predic+ons.
Here’s my radical future…
One reason we give so much to political campaigns is because we know they can make a big difference in our life for the next four years.
If theaters can better measure (and show!) their impact, and use that data to inform fundraising models —> big new revenue stream.
We’re in a rare spot in the nonprofit field of relying so little on contributed revenue. 7 nonprofits received more than a billion dollars last year. Arts orgs were only 4 of the top 100 largest charities (https://
www.forbes.com/sites/williampbarrett/2016/12/14/the-largest-u-s-charities-for-2016/#70405dea4abb)
If your theatre was using outcomes-based predictions to improve your performance, what new opportunities could that lead to?
30. Trend 3:
Big data.
New revenue opportuni+es arise when
algorithms can pinpoint the needle in
the big(ish) data haystack.
There aren’t a ton of places where there is truly big data in the theatre world, at least within a single organization (> 1M rows in Excel). So we’re really talking about medium-ish data.
I see opportunities for big data analysis in financial data, cultural data, and marketing data.
Once you bring lots of rows of data on a given topic into one place, algorithms can find valuable trends (that humans can’t).
31. Trend 3: Big data.
Financial data.
What if we could compare the business
decisions (and their effect) of every arts
organiza+on in the country?
There are tens of thousands of arts organizations who are all taking slightly different approaches to their business models.
This is highly structured data - revenue & expense streams, over many years.
If all of that financial data was in one place, we could do some amazing analysis of the impact of various business decisions on sustainability of organizations.
32. The National Center for Arts Research KPI Dashboard is one of the best examples of bringing a ton of financial data into an interactive dashboard to find trends across organizations.
Using data from the Cultural Data Profiles from DataArts.
mcs.smu.edu/dashboard/
33. Trend 3: Big data.
Cultural data.
What if we knew the loca+on of every
“cultural asset” in the country or in your
local community?
Every artist, artwork, and arts organization is a data point that could be captured and mapped.
34. The cultural asset mapping that Alliance for California Traditional Arts has been doing with partners like Arts for LA is really exciting.
This map shows the dance & theatre companies along LA’s Red & Purple metro lines. Now I can start to talk about the intersection of transportation & the arts, how important access to public transportation is
to diversifying audiences in some communities.
culturemapla.org/#
35. Trend 3: Big data.
Marke?ng data.
What if we knew the “purchase pa9erns”
of every person in a community?
Political campaigns have voter profiles of every voter in a particular district. Information about their demographics, their propensity to vote, their party affiliation (equivalent to genre interest), whether they vote
in local & national elections (equivalent to our cross-organization ticket purchases).
36. In the UK, the Audience Finder is a national audience database contributed to by hundreds of arts organizations - that includes user-level analysis of every ticket transaction, fundraising transaction, and
audience survey response, all tied to unique identifiers (like an email address). Participants can’t see actual names, but can better understand specific communities.
https://audiencefinder.org/
DataArts Demographic Initiative is working on a similar initiative on a smaller scale in Houston.
culturaldata.org/houston
37. Trend 3: Big data.
Radical future.
The Washington Post model of
disrup+on: Deeply integrate
measurement and technology into a
powerhouse content producer to turn
around a failing business model.
Here’s my radical future.
If you had access to the financial data of thousands of theatre in the country, millions of artists, or the audience profiles of 300 million Americans, how could that data change what your organization could do?
Resource: fortune.com/2017/03/13/washington-post-arc/
38. Trend 4:
Combined data
sets.
Combining data sets creates new
markets and/or meaning.
You can consolidate a bunch of similar data sets into one database to create a new market (Uber did this for rider-data).
After you combine data sets, the ability to find trends in the data can create new opportunities.
Sometimes new meaning is found when you merge arts data sets with data sets from outside of the arts.
39. Trend 4: Combined data sets.
Consolida?on.
Consolida+ng similar data from many
organiza+ons into one loca+on can
create new markets.
London Theatre Direct API puts tickets from 150+ theatre venues in one place. https://developer.londontheatredirect.com/
Data doesn’t have to be numbers - the New Play Exchange puts more than 10,000 scripts by living writers in one location. nnpn.org/
The value of these data sets arise when their numbers are large enough to create entirely new marketplaces.
Other examples:
• Art Tracks structured collections data: www.workergnome.com/work/art-tracks/
• Atom movie ticketing app: https://www.nytimes.com/2016/12/04/business/media/movie-ticketing-start-up-atom-tickets-hopes-to-fill-theaters.html?_r=0
• ArtNet auction pricing: www.artnet.com/about/aboutindex.asp?F=1
40. Fractured Atlas’ Space Finder consolidates information about available (and needed) space into one location. Now up & running across 20 different cities or states. Includes everything from parks to civic
building, bookstores w/ reading rooms not in use. All these locations are just data points. The value comes from having them in one place, easily searchable.
Created nearly $1 million in rental revenue.
Opening soon in Portland.
https://www.fracturedatlas.org/site/spacefinder
41. Trend 4: Combined data sets.
Benchmarking.
Compare data across similar
organiza+ons to benchmark, find trends
and opportuni+es.
Sometimes you want to pick & choose from many of the same data points to find the one you like. Other times you want to compare 1 data point to all the other similar data points in a set.
That ability to find a trend or comparison creates a new opportunity.
42. National Centre for Arts Research - Arts Vibrancy Index ranks more than 900 communities across the country, examining the level of supply of art, demand for art, and government support for the arts in each
city.
“The index helps us understand what factors contribute to making an urban area artistically vibrant and culturally rich”
https://sites.smu.edu/meadows/heatmap/index.html
Links:
Theatre Facts: www.tcg.org/ResearchAndResources/TheatreFacts.aspx
Orchestra Facts https://americanorchestras.org/knowledge-research-innovation/orchestra-facts-2006-2014.html
PACStats: ams-online.com/pacstats/
43. Trend 4: Combined data sets.
Unique sources.
Combine unique non-arts data sets with
tradi+onal arts data sets to create new
opportuni+es.
City governments are at the forefront of combining unique data sources from lots of different sectors into one location and seeing what new opportunities arise.
44. City of Leeds UK created an open data dashboard that combines more than 700 data sets from across the city, including nearly 50 arts & culture data sets. https://datamillnorth.org/
Other than just opening up public data, they’re also creating “data challenges” to find people to analyze & visualize data mashups between all of these data sets to find cross-sector opportunities.
45. Trend 4: Combined data sets.
Radical future.
The Tinder model of disrup+on: A mobile
app consolidates actor headshots and
bios, past performance cast and design
lists from every American theatre for
easy searching and nego+a+ng.
This is my radical future.
What new revenue opportunities might arise if you consolidated data across similar organizations, or non-arts organizations?
46. Trend 5:
New collec?on
methods.
Technology enables new data
collec+on methods, which gives rise to
valuable new data sets.
With new technology, we have the ability to collect geospatial data biometric data about visitors or artists. We can use machine learning to match patterns between images and the real thing. All of these new
sets of data create new opportunities.
47. Trend 5: New collec0on methods.
Geospa?al data.
Knowing where visitors are at any point
in +me creates opportuni+es to serve
them be9er.
Geospatial data can be gathered from mobile phones that have wifi turned on (pinging cell towers) or by bluetooth connections (using beacons). This can pinpoint location to within a foot or so.
Other examples:
• National Gallery’s wifi tracking of visitors to inform exhibit placement, duration, and marketing. www.gizmodo.co.uk/2017/04/exclusive-heres-what-museums-learn-by-tracking-your-phone/
• The Art Institute of Chicago leveraged its 300-beacon network (activated when visitors connect to WiFi) to increase paid attendance from $14.8 million in 2015 to a projected $19.9 million the following fiscal
year. https://www.artsy.net/article/artsy-editorial-can-big-data-make-for-better-exhibitions
48. The Brooklyn Museum uses beacons to help power its ASK mobile app. Curators are responding to text messages by visitors. Team knows where visitors are based on beacons. Average depth of
conversation = 14 messages.
Knowing the geospatial locations provided deeper interactions w/ the art.
https://www.artsy.net/article/artsy-editorial-can-big-data-make-for-better-exhibitions
49. Trend 5: New collec0on methods.
Biometric data.
Use biometric data to understand
physiological reac+ons to events.
In the same way watches & fitness apps measure heart rate, perspiration, breath rate to judge our response to physical activity, so can we use similar data to judge our response to aesthetic inputs.
Other examples:
• Dolby biometric sensors to evaluate films: https://www.theverge.com/2017/3/19/14949798/dolby-labs-biosensors-eeg-brain-heart-rate-movie-tv-reactions
• NIDMS dance injury prevention: https://www.nidms.co.uk/research
50. Body Metrics exhibit at the San Jose Museum of Innovation fitted visitors with mobile devices to monitor their heart rates, brain waves, muscle tension and more as they browsed the museum. Carlson said,
“the point is how we can use Big Data technology to take small steps to change our behaviors and improve our mental and physical well-being. Like, by knowing that I get tense at certain moments, I can be
reminded to take deep breaths and calm myself. Or if the system sees I’m feeling exhausted, it will suggest I need to hydrate.”
www.mercurynews.com/2014/10/29/new-body-metrics-exhibit-opens-at-the-tech-its-all-about-you/
51. Trend 5: New collec0on methods.
Image Recogni?on.
Use pa9ern recogni+on to iden+fy
“difficult to recognize” data sets.
Machine learning can find patterns in data sets after being “taught”
52. Magnus is the Shazam for Art. Take a picture of a painting, and instantly get the artist, title, and price.
www.magnus.net/
Ensemble@Yale uses image recognition to transcribe 90 years of archived program history.
ensemble.yale.edu/#/
53. Trend 5: New collec0on methods.
Radical future.
The Google Ads model of disrup+on:
Google knows what you search for.
Theatres know how you feel.
This is my radical future.
What if you had access to geospatial or biometric data about visitors? What if image recognition would instantly identify & classify whatever you saw?
54. Trend 6:
Consumer access.
Giving a user access to relevant data
changes their behavior.
Access might go to stakeholders, to users, or about other peers.
55. Trend 6: Consumer access.
Stakeholder access.
Provide (near) real +me transparency to
community stakeholders.
More stakeholders are requesting or demanding access to theatres’ data - whether those are institutional or individual donors, city governments, boards of directors, or audiences.
56. Dallas Museum of Art dashboard. Dashboards open up data to anyone and everyone in an organization to make decisions and to signal that data is important. https://dashboard.dma.org/
Real time metrics can empower stakeholders
Other dashboards:
Smithsonian dashboard.si.edu/
Carnegie dashboard.carnegiemuseums.org/
MOMA https://medium.com/digital-moma/hey-could-you-give-me-the-numbers-on-that-again-2dd7e8060d83
Met www.metmuseum.org/blogs/digital-underground/2015/data-stories-centralized
Cleveland dash.clevelandart.org/
University College London citydashboard.org/uclmuseums/
57. Trend 6: Consumer access.
Personal access.
Provide access to users of their own data
to influence their behavior.
Like the quantified self movement has shown, giving users access to their own data can change their behavior. Access to the data becomes addictive.
58. CooperHewitt Visit Statistics page records your visit to the museum and summarizes the data in ways you might not necessarily know without the data.
https://labs.cooperhewitt.org/2016/the-visit-statistics-page/
59. Trend 6: Consumer access.
Peer access.
Provide access to how you compare to
peer (or previous) data points to
influence behavior.
Sometimes it’s not about knowing our own data, but how we contribute to the whole picture, or how we compare to our peers.
60. National Centre for Circus Arts In UK was a client I was working with on a marketing project.
They had a huge old building, ever increasing energy bills and wanted to help staff make better decisions about their energy usage.
Installed a device on their front desk that was a visual representation of their current utility meter reading. Glowed red when they were over average, green when under. Reduced costs and more sustainable.
prnewslink.net/releases/9206.html
61. Trend 6: Consumer access.
Radical future.
The Fitbit model of disrup+on: It’s cool to
be cultured.
My radical future - widespread perception change about the arts, as has happened for health.
What could change if you provided more data to your stakeholders, your visitors, or your staff?
62. Trend 7:
Data as art.
The data itself becomes art.
In some areas, data about a subject matter has become as popular as the subject matter itself. Think fantasy sports, or political polling, or health apps.
We’re going to look how data is being used not only for its objective value but for its aesthetic value.
63. Trend 7: Data as art.
Feedback loops.
Data about the ar+st informs the ar+st’s
work.
Artists are using their own data as an input into the artistic process.
64. In March of this year, a technology company wanted to demonstrate their capabilities by working with a choreographer and dancer to show off the tech.
Thus was born the Ballerina Project - this dancers EKG, acceleration, location, and sound is being recorded and displayed in real time. The dance was choreographed in time to these biometric data points.
https://www.elastic.co/blog/on-indexing-heartbeats-and-ballet-into-elasticsearch
65. Trend 7: Data as art.
Interpreta?on.
Data about the art becomes art in itself.
Data visualizations are becoming works of art in their own right.
66. Suns Explorer came out of the Harvard Art Museum. It’s an interactive infographic capturing the color of art works in the museum’s collection. This is a still shot, but the real thing is a constantly moving,
refreshing video about art, that makes art.
https://www.bostonglobe.com/lifestyle/2017/04/13/when-big-data-meets-art-appreciation/HqeuVGv9qdm2PGJAeYAuZK/story.html
67. Trend 7: Data as art.
Data transformed.
Art is created from data sets.
Data is being used for purely aesthetic reasons. It creates the artwork.
68. The World Science Festival hosted the Umbrella Project last week, a collaboration between MIT’s Distributed Robotics Laboratory and performance arts organization Pilobolus Dance Co, invites visitors to
take part in a unique performance piece. Armed with controllable LED-lighted umbrellas, participants will form colorful constellations and patterns in response to questions asked.
69. Trend 7: Data as art.
Radical future.
The Hamilton model of disrup+on: An
unexpectedly smash hit teaches the
world about a (for some) boring topic
like data.
Here’s my radical future.
What if it’s not just data that helps the arts, but the arts that can help transform the data field.
70. But what can I
do with this data?
So we made it through 7 trends, but how are those impacting arts organizations? Let’s recap.
71. Single 8ckets.
Individual
dona8ons.
Ins8tu8onal
giving.
Board giving.
Other
contribu8ons.
Other earned
revenue.
Subscrip8on
8ckets.
Big data.
Personaliza8on.
Outcomes-based predic8ons.
Consumer access.
Combined data sets.
New collec8on methods.
Data as art.
Revenue
+ trends.
If what you care about is increasing your single ticket sales or your individual donations, I would focus on using data for personalization & outcomes-based prediction.
If you’re more interested in opening up new, non-traditional revenue opportunities, focus on combining data sets, using new collection methods, or experimenting with data as art.
72. Ar8st payroll.
Produc8on
costs.
Opera8ons.
Marke8ng &
development.
Administra8ve
payroll.
Produc8on
payroll.
Expenses
+ trends.
Personaliza8on.
Outcomes-based predic8ons.
Big data.
Combined data sets.
Consumer access.
Outcomes-based predic8ons.
New collec8on methods.
Data as art.
Combined data sets.
If you want to think about increasing the efficiency of the cost of your building & general operations, think about combining data sets and giving consumers access to their data.
73. 7 elements of a data
culture.
1. People.
2. Skills.
3. Process and values.
4. Tools and infrastructure.
5. Access.
6. Ethics.
7. Funding.
7 critical elements of a strong data culture within an arts organization.
Take an inventory of what you have and what you need.
Roughly in order of how much organizations in the early stages need to be thinking about these things.
Resource: Data Driven Museums https://zenodo.org/record/321812#.WTj-8YnytE6
74. Element 1:
People.
Data roles can be temporary
residencies from other organiza+ons,
permanent solo staff members, or
en+re departments/labs.
Temporary → Arts Data Impact residency placed a Data Scientist in 4 different UK arts organizations and documented their work over 2 years.
• Also: Uptake Fellows, LA DCA Design Lab, V&A last month posted a 6-month residency
Permanent staff: Data Scientist, British Museum (Alice Daish).
“Open Day” to diversify tech applicants to Natural History Museum
• Executive Director of Analytics, Art Institute of Chicago (Matthew W. Norris)
• Data Analyst, Museum of Fine Arts (job description)
• Digital Data Analyst, British Museum (job description)
• Associate Director, Evaluation and Visitor Insights, OMCA Johanna Jones
Team: SF MoMA labs is a good example of a team of digital, technology, analysts, and other staff. Frequent blogging.
• More labs: Cooper Hewitt, NYPL, Science Museum UK, Carnegie, IMA, BKM Tech Lab
Links:
www.artsprofessional.co.uk/magazine/280/faces/data-scientists-join-nt-barbican-and-eno
https://datafellows.uptake.com/
dcaredesign.org/lab/
www.vam.ac.uk/blog/artists-residence-va/new-open-call-for-nominations-applications-vari-embedded-residency-for-practitioners-working-with-data
https://careers.nhm.ac.uk/templates/CIPHR/jobdetail_273.aspx
labs.aam-us.org/blog/the-power-of-applied-data-for-museums/
www.mfa.org/employment/data-analyst
https://www.linkedin.com/jobs/view/149275169/
https://www.sfmoma.org/series/sfmoma-lab/
https://labs.cooperhewitt.org/
https://www.nypl.org/collections/labs
https://lab.sciencemuseum.org.uk/
https://studio.carnegiemuseums.org/
lab.imamuseum.org/
https://www.brooklynmuseum.org/community/blogosphere/
75. Element 2:
Skills.
Finding a problem to solve is the best
way to learn data skills (sta+s+cal,
technical, systems theory, visual
design, communica+ons).
Problem solving is more productive than abstract. 5 major skills a data team needs.
Conferences, hack events, and online resources can also help you build capacity.
• ArtsDatathon event
• Other events: Creative Data Club, TCG Measuring Up, DataKind, SciHack + MOMA Datathon
• Data on Purpose conference (Stanford, February)
• Other conferences: CSVConf, Open Analytics, DAA, Eyeo Festival, Museums & the Web, Museum Computer Network, MuseumNext, Museum Computer Group, CultureGeek, NAMP
• Connecting the Data - TRG & DataArts online course
• 65 Data Science Skill Building online classes
• CulturalDigital email newsletter
• Other online resources: 125 Data People on Twitter, DataCamp
Links:
https://artsdatathon.org/
www.ssirdata.org/
www.trgarts.com/TRGInsights/Article/tabid/147/ArticleId/432/Online-course-Connecting-the-Dots.aspx
https://elitedatascience.com/data-science-resources
us4.campaign-archive2.com/home/?u=a5109b3260499a3d8b78a4780&id=f5c318bb03
https://twitter.com/devonvsmith/lists/data-folks/members
https://www.datacamp.com
76. Element 3:
Process and values.
Embed data (culture) throughout the
organiza+on.
• Open the data/knowledge to everyone who wants it
• IOW: create knowledge sharing opportunities, break down data silos within the organization (MCN video1 & video2)
• Make data analysis selective, routine and pervasive
• IOW: Don’t try to measure everything, don’t only do it annually, don’t relegate it only to the marketing department
• conduct experiments and pilots (in opera sector; in journalism)
• IOW: Get comfortable with incremental change and risk taking
• Spend equal time (25% each): gathering, cleaning, analyzing, sharing data
• We spend a lot of time on the 1st thing now, a little time on the 3rd.
Links:
https://www.youtube.com/watch?v=2taI9hISEN4
https://medium.com/digital-moma/what-does-data-have-to-do-with-it-5f4c1d95da14
https://static1.squarespace.com/static/51d98be2e4b05a25fc200cbc/t/56cd1cd986db43bd3f1dac17/1456282842159/Kevin+final+paper.pdf
https://journalistsresource.org/syllabi/data-journalism-visualization-mapping-ethics-syllabus
77. Element 4:
Tools and
infrastructure.
Different data tools allow you to
analyze data, visualize findings, and
automate the analysis process.
These tools are readily available, cheap/free, and reasonably easy to learn.
• Arts specific: Dexibit, Datorama/Capacity Interactive built tool announced in Mar ‘17
• Beginner: Google Explore Sheets, Slack Statsbot
• Intermediate: Tableau, Google Data Studio, Plotly, InfoActive, UnionMetrics, SparkWise, SimplyMeasured, ChartBeat, ChartBuilder, Trifacta, DataBasics.io
• Advanced: British Museum’s use of r 17-min video
Links:
dexibit.com/
www.marketwired.com/press-release/capacity-interactive-puts-data-work-arts-with-datoramas-marketing-intelligence-solution-2206155.htm
https://blog.google/products/docs/explore-docs-sheets-and-slides/
https://statsbot.co/slack
labs.aam-us.org/blog/becoming-a-data-startup-part-iii-goals/
https://channel9.msdn.com/Events/useR-international-R-User-conference/useR2016/Transforming-a-museum-to-be-data-driven-using-R
78. Element 5:
Access.
Provide access to your data via
internal data warehouses, open APIs,
and other data portals.
• Most “data warehouses” are proprietary, but some are public
• Github repositories: Warhol Museum, Met, Carnegie, Cooper Hewitt, Tate, MOMA
• Cooper Hewitt API - opens up an institutions data for others to use and build on top of (new products)
• Other APIs: DMA, OpenCultuurData, PowerHouse, SF Moma, SciMuseum UK
• Arts Datathon data portal - collect data across institutions in a less structured way
• Other data portals: UK, data.gov, DataSF, datausa.io, Sustain Arts, Tempe
Links:
https://medium.com/@caw_/open-data-from-tms-for-all-5c68b5adcad6
https://collection.cooperhewitt.org/api/
museum-api.pbworks.com/w/page/21933420/Museum APIs
https://artsdatathon.org/data/datasets/
jamesdoeser.com/data-portals-in-art-and-culture/
data.gov
https://datasf.org/opendata/
datausa.io
southeastmichigan.sustainarts.org/#/detroit
kjzz.org/content/481186/potholes-police-calls-and-more-found-tempe’s-open-data-portal
79. Element 6:
Ethics.
Staff and cons+tuent data is at risk
without privacy and security protocols.
Your organiza+on’s integrity is at risk
without consent and transparency.
Privacy, security, and ethics are critical to any data team, but a nascent topic in field.
• San Antonio musician data breach
• GDPR - international data protection law
• Data ethics cheat sheet of resources
• Also beware of researcher bias.
• visualization bias, Machine bias in jail sentencing; Machine bias in medical reimbursements
Links:
www.expressnews.com/business/local/article/Data-breach-hits-San-Antonio-Symphony-employees-10931740.php
https://www.civilsociety.co.uk/news/free-guide-to-gdpr-and-data-protection-for-charities-published-today.html
https://hackernoon.com/the-big-data-ethics-cheat-sheet-34999f751529
https://flowingdata.com/2017/02/09/how-to-spot-visualization-lies/
https://www.propublica.org/article/machine-bias-risk-assessments-in-criminal-sentencing
https://www.aclu.org/blog/free-future/pitfalls-artificial-intelligence-decisionmaking-highlighted-idaho-aclu-case
80. Element 7:
Funding.
Funding to invest in data can come
from opera+ons, grants or get re-
invested from savings by data
ini+a+ves.
I believe by-and-large, funding is and should be the least important issue in this line up of 7 elements. Don’t wait on funding to be available, work with what you have.
• NESTA UK - government support of data in the arts (UK), NEA Research grants
• Transforming Lives Thru Arts: social impact measurement for Dallas arts orgs, IMLS Digging into Data
• UK Nudge Unit earns 20x back investment in 2 years
Links:
www.nesta.org.uk/
www.taca-arts.org/news/2017_TACA_announcement.php
https://diggingintodata.org/awards/2016
www.socialsciencespace.com/2015/11/welcoming-the-american-nudge-unit/
81. Q&A.
#TCG17
@devonvsmith
slideshare.net/devonvsmith
www.measurecrea;ve.com
Does data have any negative impacts on the arts?
What about how data will affect other industries, which will have spillover effects for arts?
Myths holding us back
• Data about past sales leads to bad (artistically) new productions (counter: netflix)
• Data only raises ticket prices (counter: medical billing, lyft)
• Data leads to segregated markets (counter: increase college application diversity)
• It’s hard & expensive to collect information about our communities (counter: political campaigns do it quickly, aggregate many sources)
• Our data is too complicated, too many sources, too diffuse (counter: wind farms → weather, deforestation, tidal, space, finance/markets)
82. 7 elements of a data
culture.
1. People.
2. Skills.
3. Process and values.
4. Tools and infrastructure.
5. Access.
6. Ethics.
7. Funding.
7 trends where data is
transforming the arts.
1. Personaliza8on.
2. Outcomes.
3. Big data.
4. Combined data sets.
5. New collec8on methods.
6. Consumer access.
7. Data as art.