Blog

Participting in the Data Science for Social Good Incubator Program

We’ve been selected to participate in the Data Science for Social Good summer incubator program!  It’s a really great opportunity for us to work with smart graduate students and faculty at the UW eScience Institute, while sitting side by side with three other selected projects leveraging data science for social good in urban environments.

eScience_DSSG

Here’s the project description:

Assessing Community Well-being through Open Data and Social Media

With this project we will be creating neighborhood community report pages in the context of a hyperlocal, crowd-sourced community network. Our direct, “socially good” objective is to help neighborhood communities better understand the factors that impact community well-being, and how they as a neighborhood compare with other neighborhoods on these factors, to help them set the agenda for what to prioritize in promoting their well-being. A key aspect of this project is to explore novel ways to leverage diverse social media and open data sources to dynamically assess community-level well-being, in order to a) enable early identification of emerging social issues warranting a collective response, and to b) automatically identify and recommend the local community hubs best positioned to coordinate a community response. While the tools are intended to be general purpose, through the summer we will be targeting two more underserved neighborhoods in King County (the International District and the Central District). Specific project activities include:

1) Collecting and processing diverse hyperlocal social media (e.g., Twitter, Instagram, Foursquare) and open data sources (e.g., Census data, Crime data, building permits, map data) to develop community well-being measures, which may include sentiment/content analysis, social network analysis, geo-spatial analysis of hyperlocal business activity, and social media activity metrics;

2) Algorithmically integrating these metrics to develop a summary measurement model of overall “community well-being” and individual/local business “community hubbiness”;

3) Representing these metrics to end users (neighborhood community members) in neighborhood report pages, which may include overview visualizations that represent neighborhood well-being across neighborhoods; and

4) Throughout the process, actively work with one or two underserved neighborhoods (International District, Central District) to engage in community participatory design.

This works builds on prior work while our lead investigator Shelly Farnham was as Microsoft Research, exploring the use of social media to help hyperlocal communities stay informed. Related references:

Whoo.ly: Facilitating Information Seeking For Hyperlocal Communities Using Social Media. Yuheng Hu, Shelly Farnham, Andrés Monroy-Hernández. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (CHI) 2013

Neighborhood Community Well-being and Social Media, Shelly Farnham, Michal Lahav, Emma Spiro, and Andres Monroy-Hernandez. A recent, as yet unpublished paper exploring the relationship between social media usage and hyperlocal community well-being.

Recipient of Microsoft Azure for Research Award

I’m excited to announce we are a recipient of the Microsoft Azure for Research Award, which includes large allocations of Azure storage and compute resources for a period of a year.

While at the iConference in March, I saw a demo of Azure’s machine learning for research functionality and found out about the award program.  We use Azure for our database backend, and given the level of data analytics we will be performing in the next year, this is a great program for us.  They’ve been developing some pretty sophisticated machine learning tools in the past few years, including integration of R, and a graphical predictive analytics tool.  We are excited to try them out.

Thanks Microsoft Azure!

2248_MSR-MachineLearning-blog-final_png-496x0

Innovative Community of Practice and Civic Tech in the Northwest

I gave a couple of talks in April discussing how to foster innovative communities of practice  (one at the Cultural Congress hosted by the Washington State Arts Alliance, and one for the e-gov subcommittee of the CTTAB).

Here are the slides as a PDF.

Presentation highlights:

Due to the ubiquity of social media, there’s been a paradigm shift in how group’s of people work together, which also affects how organizations (government, NGO, or companies) engage with the public.

There are several models for citizen engagement via social media, the one we at Third Place Technologies seek to explore is how to best provide environments that empower members of communities to connect with each other — helping communities help themselves.

modelsofengagementviasocialmedia

 

We are working on a couple of prototyping projects that are focused on helping communities connect, including Brink NW, a crowd-sourced community network of people, organizations, events projects, optimized for interdisciplinary, cross-organizational collaboration.

As a part of this effort, we are focusing on three communities to start — the art & tech community in the Northwest, the civic tech community in the Northwest, and the International District as a neighborhood community.  A key aspect of this project is to engage in data-driven, socially intelligent design — that is, to develop a community networking tool that best maps onto the existing structure of real-world communities of practice.  Consequently we’ve been actively engaged in acquiring a lot of information about these communities, including what we are calling a “community mapping survey”.

We recently distributed one of these community mapping survey’s to the civic tech community at the Open Data Day in February, and presented some key findings from this survey (see slides 21-32).

In asking participants to articulate the community’s goals, they expressed the strongest interest in having civic tech exhibitions/show and tells, and then educational workshops.

Civic Tech Community Goals

Civic Tech Community Goals

It was clear from the survey, that as a new, emerging community of practice, people had some difficulty connecting with others — knowing who’s who, and what projects were active, and whey they could plug in, so to speak, with their skill sets.

Welcome to Our New Board of Directors

Perhaps the most important task in founding a new non-profit organization is forming a board of directors that is not only excited about the core mission, but will also provide the expert advice, ethical oversight, and social connections required to assure the organization’s success.

Third Place Technologies was incorporated as a state non-profit with an interim board in October of 2014, and we have since been engaged in the process of finding and recruiting more board members, seeking a diversity of backgrounds and skills to assure we have experienced research scientists, technologists, nonprofit administrators, community organizers, and designers at the table.  We held our first formal meeting of the board on March 31st, where we adopted our bylaws, elected our board officers, and approved our 1023 application for 501(c)(3) tax exempt status.

Welcome to our new Board of Directors, including Bob Mason (President), Joe McCarthy (Vice-President), Dee Christoff (Secretary-Treasurer), Jeff Larson, Genevieve Tremblay, Michal Lahav, and Seth Vincent, and thanks for your commitment to Third Place Technologies’ core mission of creating innovative technologies that foster community empowerment and well-being.

As of this past week, our documents are all signed and in the mail, which is a great cause for celebration.

Designing for the Perils of Public Voice in the Age of Social Media

On February 26th I gave a talk at the Seattle Pecha Kecha #60, Watch Me Now, Notes on a Surveillance Society.  The event brought together speakers from across the information ecosystem – including policy makers, technologists, advocates and others to discuss the complex issues surrounding privacy and surveillance in the digital world.

For this talk I discussed some of the themes that inspired the formation of Third Place Technologies — the power and perils of participating in public discourse online in the age of social media.

Here Be Dragons

Here Be Dragons

As noted by Bennett et al., (2009), the mere act of participation in public discourse online is a form of civic engagement as people share their personal stories, find people with similar interests and concerns, and then start working together to solve their collective problems.  We have consistently found in our own research (e.g., Farnham et al. 2013) that it is the use of public networks (such as Twitter, blogs) that most corresponded with civic engagement, not, the use of personal networks (such as Facebook).

However, not all voices are equally represented in public spheres online.  For example, while there are roughly an equal number of men and women with Twitter accounts, 64% of private account are women vs. 36% of men.  (See statistics on beevolve.com.)  In other words, women are not using social media to participate in public discourse as much as men are.

The story of #YesAllWomen provides a compelling example of the power and perils of online social media.  This Twitter hashtag started in response to Elliott Rodger’s killing spree in May of 2014, where he killed 7 people including himself.  To explain his murderous actions he posted a Youtube video stating “you girls have never been attracted to me, but I will punish you for it”.  This chilling event evoked a strong reaction for many women, and within days, 1.2 million posted on Twitter their stories of fear of harassment and sexual assault from men who were similarly angry for not receiving sexual attention from them, under the hashtag #YesAllWomen.  For many women, this outpouring of stories proved an empowering experience, giving a collective voice to everyday experiences that need to addressed.   Unfortunately, the stream of messages was also interspersed with threatening and antifeminist posts.  Within days, the woman who had created the original  #YesAllWomen hashtag received so many threats she had to change her account to private.

In other words, she was silenced.  This is not an unusual occurrence, many are silenced from participating in the public sphere through real or anticipated threats of intimidation.   For example, Anita Sarkeesian, a feminist influential who was speaking out against the violence against women in video games, receives hundreds of abusive Tweets every week including death and rape threats.  Just recently she received a mass shooting threat for a scheduled talk at Utah State, which she felt impelled to cancel because there was not sufficient security.  (Learn more here.)  She also was silenced, though she assures us only temporarily.

Outspoken feminists experiencing systematic intimidation are just one example of the perils of participating in the public sphere — there are many, many others; ranging from a gay teacher posting marriage announcements on Facebook and losing his job, to an anti drug cartel advocate in Reynosa, Mexico who was murdered for her Tweets reporting on their crimes.

Given the importance of being able to engage in the public sphere online, how can we help design third places technologies to assure the benefits of having a public voice online are experienced without harm by all?

As I mentioned in this talk, it is important to recognize that online environments can be hostile, much like work environments can be hostile, characterized by the cumulative effect of conduct in social media so severe and pervasive it interferes with a citizen’s ability to engage in public discourse.     Systematic acts of intimidation through threats of violence are not free speech — they are the opposite of free speech, specifically designed to prevent a citizen from engaging in public discourse.  As such, it is important to develop systems for civic discourse that adopt a zero tolerance policy toward such acts of intimidation.

Similarly, in cases where disclosing civic opinions might negatively impact a person’s well-being — such as job loss, public shaming, physical harm, or even loss of life — it’s extremely important to ensure participants have the right tools for privacy and identity segmentation.  For example, I should be able to actively engage in authentic discourse around my political beliefs without my persona as an engaged citizen being tied to my persona as an employee or a family member.  In face-to-face situations we can handle identity segmentation quite effectively simply by changing places (home, work, town hall), but have difficulty doing so in online systems that tend to collapse our social contexts — and identities — together.

Tools such as Twitter and Reddit are more appropriate for civic discourse than tools such as Facebook, because they enable pseudonymity.  This is an identity separate from the “real” you, so that participation in public discourse online does not easily lead to personal harm offline.  Yet, this online identity is persistent, and may develop its own reputation over time, which increases accountability and socially appropriate behavior.  In this case, you might think of Twitter and Reddit as playing the role of identity brokers — a trusted, third party environment where individuals may disclose personal information.

That said, these systems have far to go before they are safe environments for people to express their public voice.  Going forward, it is important to advocate social policy fostering a culture of respect online, recognizing that threats and bullying online can lead to real harm offline.  Similarly there needs to be the appropriate legal adaptations balancing free speech and the principle of no harm.  However as technologists, there is also much we can do in our socio-technical system designs to mitigate real threats to well-being.   How we may design new technologies to ensure equitable participation and voice in the public sphere is a hard problem to solve, and an important theme of our work here at Third Place Technologies.

Here are the slides from the talk: Designing for the Perils of Public Voice in the Age of Social Media

References:

Bennett, W. L., Wells, C., and Freelorn, D. 2009. Communicating citizenship online: models of civic learning in the youth web sphere. Civic Learning Online Project.

Farnham, S.D., Keyes, D., Yuki, V., and Tugwell, C.  (2013).  Modeling youth civic engagement in the new world of networked publics.  AAAI International Conference on Weblogs and Social Media (ICWSM).

 

2015 iConference Social Media Expo — Open Government Data and Social Media

opengovntYesterday we held our third Social Media Expo at the 2015 iConference, sponsored by FUSE Labs at Microsoft Research.  This has been a great way for us to encourage interdisciplinary student teams to explore emerging themes in the field that are also of particular interest to FUSE Labs.

The 2015 Theme:  Open Government Data and Social Media

This year’s challenge was to “explore new and exciting ways to foster societal awareness and conversation at the intersection of open data and social media”.  In 2009 the white house sent out an Open Government Directive, requiring government agencies to adopt the principles transparency, participation, and collaboration in part by publishing government data online in open formats.  Five years later there is an amazing plethora of information freely available, enabling people to “conduct research, develop web and mobile applications, design data visualizations, and more”. (see http://www.data.gov).  At the same time, an analysis of social media data helps us understand and reflect on historical and real-time aggregate behavior, allowing both organizations and individuals to gain a deeper understanding of their society. These two trends create an opportunity for a kind of large scale conversation – as governing agencies share their data with their citizenry, and as the population at large engages with organizations through tools such as Twitter, blogs, and Facebook.  We encouraged teams to approach this year’s theme with an interdisciplinary combination of user research, design explorations, prototyping, real world deployments with usage analysis, or community engagement.

The winning teams:

We were very pleased with the quality of the submissions this year, and we to thank the five winning teams who received the grant for all the work that went into their projects and their presentations at the Social Media Expo.   The best project award went to the University of California, Irvine’s team, for their project “ Racial Violence Archive: Public Information System on Incidents of Violence during the Civil Rights Period”.

eMigrate: Aggregating Government Open Data for Enhanced Job Category Selection in Support of Immigration Applications

School: University of Toronto: Faculty of Information
Team: Eva Hourihan Jansen, Jenna Jacobson, Gabby Resch
Faculty Sponsor: Rhonda McEwen

The Police Officer Involved Homicides Database Project

School: University of California, Los Angeles: Department of Information Studies
Team: Morgan Currie, Brittany Paris, Irene Pasquetto, Jennifer Pierre, Ashley E. Sands
Faculty Sponsor: Leah Lievrouw

Racial Violence Archive: Public Information System on Incidents of Violence during the Civil Rights Period

School: University of California, Irvine: The Donald Bren School of Information and Computer Sciences
Team: Hosub Lee, Michael Bellato, Sowmya Jain, Fernando Spanghero, Roeland Singer-heinze, Ya-Wen Lin, Sunakshi Gupta, Geoff Ward
Faculty Sponsors: Alfred Kobsa and Geoff Ward

Society Key: Integrating Social Media Data with Governmental Open Data to Encourage Community Wellbeing

School: Rutgers, The State University of New Jersey: School of Communication and Information
Team: Ziad Matni, Jennifer Sonne, Dongho Choi
Faculty Sponsor: Chirag Shah

TransparencyScience. Return on research investment, where do the funds go?

School: Polytechnic University of Valencia: School of Informatics
Team: Lidia Contreras, Cristina I. Font, Paulina Morillo, Diego Vallejo
Faculty Sponsor: Antonia Ferrer Sapena

Emerging Themes and Discussion:

Following the presentations we had a brief discussion.  Common threads across the talks and discussion include:

  • Filing the gaps.  As noted particularly for the Police Officer Involved Homicides project, there is often a gap across different sources of data, and social media may be used to help fill the gaps to increase accuracy and interpretability.  Government data can also sometimes be out of date, and social media may be used to help bring it up to date (e.g., classification of job categories for the eMigrate project).

  • Translating data for end users.  Government data as shared in its raw form can be fairly opaque to the average person.  A key problem that can be addressed by information technologists is transforming it through simplification, explanations, and visualizations, so that it is actionable for everyday citizens.

  • Community Engagement.  Student teams generally found that people were very interested in engaging around the data, and further found that a hackathon format was quite successful vehicle for enabling deep engagement and community connections with like-minded others.  As researchers/information technologists, they found they tended to take on a teaching role, helping people to engage with the data.  Many have low levels of self-efficacy around engaging with open data and technology, and require some hand-holding.

  • Developing the community of practice.  Researchers, technologists, and “digital journalists” need to develop best practices for dealing with ethical issues, increasing engagement and end-user confidence, and assuring projects are not exploitative.

  • Social media facilitates participatory.  Across the student projects, we observed social media and social media design patterns could play a meaningful role in helping the new open data environment become more participatory and collaborative.

Thanks again to our student teams for their hard work, their faculty sponsors for providing guidance, the organizing committee for developing the theme and writing reviews, and  the iConference for generously hosting the Social Media Expo.

Please learn more about the team project by reading their abstracts or watching the video submission on Youtube.

Designing for Hack the Commute

Local civic tech hackathons and uncoferences are a great way to get to know a community even as it is gelling into one.  I went to Hack the Commute, hosted by the City of Seattle and the Washington State Department of Transportation, and was very happy to see a number of familiar faces there.  It was a great event, and very well run — a good role model for future hackathons.

As someone who can wear multiple hats I am never quite what role I will fall into at these events — research scientist, data analyst, developer, designer, artist — however by some intuition I brought my Mac to this hackathon (loaded up with Illustrator and Photoshop) and joined the Dokoji team as the designer for a day.  They are developing an app that “turns your conversation in to smart decisions”.     It was a very interesting problem — how might we integrate different sources of open data relevant to the moment-to-moment context of a text conversation, as people seek to converge in place and time?

Here are some of my design mockups:

Design mockups exploring integrating contextually relevant open data and social media into text messaging conversation.

Design mockups exploring integrating contextually relevant open data and social media into text messaging conversation.

The full slide deck for our short wrap-up presentation is posted on github here.  To quote their summary:

“For Hack the Commute, we integrated data sets that would allow people to explore green transportation options, so once a group of people had picked a destination, they would be presented with 3 options: Find the closest bus using the OneBusAway data set; find the closest bike rental location using the Pronto data set; and a carpool option (future implementation). On the back end, we also integrated impact data from Washington State Department of Transportation. “

Thanks to the Dokoji team, it was great fun hacking the commute with you.

Apolis: Our Third Place Heatmap Project at HackHousing

This past weekend I attended “HackHousing”, a hackathon hosted by Zillow in coordination with the White House, U.S. Department of Commerce and theDepartment of Housing and Urban Development and in partnership with the Department of Computer Science & Engineering at the University of Washington.

It was a great hackathon, very well attended and facilitated by Chris Metcalf from Socrata.

I was very lucky with my impromptu team (Saba, Daniel, Elizabeth, and Jim), who were excited to do a project exploring how can we indicate that a neighborhood has a thriving community based on open data. Building on a long history of research showing that thriving communities are more likely to grow when there are “third places” in a centralized location, we created a heatmap quantifying the number of third places in a neighborhood and the neighborhood’s community “hubbiness”.

heatmpap

Apolis, with heatmap indicating community places, and bullseyes indicating neighborhood community centroid & density

 

In addition to the heatmap, we found the centroid and density of each neighborhood’s hubs and represented them with little bullseye’s on the map (larger bullseye = more third place density) so people would know where they should live.

We used the City of Seattle’s neighborhood map data to find places such as libraries, parks, community centers, and schools.

We also used the Location Affordability Index open data provided by HUD for the event to help people filter over the map.

The code is shared on github here.

Upcoming Seattle Open Data Day

I’ve been helping Seth Vincent (Code for Seattle) organize Seattle Open Data Day with Will Scott at the University of Washington.

I strongly encourage you all to go to Seattle Open Data Day, an unconference on open data and civic technology.  Held at the UW Computer Science Department on February 21, the goal is to continue the conversation on civic technology issues around Seattle.

Some of the topics we hope to discuss are:

  • Defining what we as a city/region want from open data in the future, and how the government & companies can help.
  • Making it easy for everyone to participate in open data and civic technology.
  • Where open data should be used to make and improve decisions.

More information here:  http://codeforseattle.org/open-data-day

Sign up through the Code for Seattle meetup group here: http://www.meetup.com/Code-for-Seattle/events/220166485

Consider leading a session! http://codeforseattle.org/open-data-day#lead-session

Draft of paper Neighborhood Community Well-being and Social Media

twittergraph

In neighborhoods that actively used Twitter, Twitter message positivity correlated with self-reported neighborhood community well-being.

With the transition leaving Microsoft Research and starting up Third Place Technologies, I have not had much time for revising this paper, but still wanted to throw it out for people to read if they are interested.  This line of research has hugely influenced my own thinking in this space, including my decision to venture in to the non-profit sector to more proactively apply some of the lessons learned to real world problems.

Here is the PDF

Title: Neighborhood Community Well-being and Social Media

Shelly D. Farnham Microsoft Research (formerly), Michal Lahav, Living Research Labs Seattle, Andres Monroy-Hernandez, Microsoft Research, and Emma Spiro, University of Washington

Abstract: In the following study we adopt a multi-method approach to examine whether the growing use of social media as a channel for hyper-local conversation may provide meaningful insights into the well-being of neighborhood communities. First, through interviews and a questionnaire with 174 residents of 26 neighborhoods we explore what are indicators of neighborhood level well-being, and what are current communication practices around the use of social media to support community well-being. Second, through an analysis of neighborhood-level Twitter messages we examine the extent to which mood and social interactivity in Twitter correspond with our neighborhood well-being indicators.  Overall, we found self-reported usage of social media positively correlated with community well-being. However, while smaller neighborhood communities had higher community well-being, they were lower in usage of social media for interacting with neighbors.  Only in larger, more urban centers characterized by younger professionals, did Twitter message mood and social interactivity correlate with well-being.

Indicators of Community Well-being Percent Mentions
Thriving local businesses 47%
Safe, low crime 33%
Community events 25%
Community resources 25%
Friendly 25%
Walkability 25%
Gathering places 24%
Social support 20%
Well-maintained 19%
Other health: mental, economic, physical 19%
People know each other 14%
Diversity (race, SES, age, families) 12%
Vibrancy — people out and about 11%
People interact/communicate 11%
Civic engagement 10%
Environmental/geographical assets 10%
Growth – embracing change 10%

“What does this community have that indicates to you that it is healthy or unhealthy?”