Vera Data 2.0 Slides and Script
0 likes121 views
This presentation analyzed data from 200,000 Causes users to understand user types. Quantitative methods like k-means clustering and decision trees identified 6 personas. Surveys and interviews with 1,466 and 65 users provided qualitative data. The results identified six main user types of Causes: Casual Activist, Self-Assured Millennial, Practical Activist, Ambitious Activist, Organized Retiree, and Tenacious Veteran Activist. Future work will engage users through tailored approaches based on their needs and identify ideal user types.
Vera Data 2.0 Slides and Script
- 1. Who are Causes users? An Exploratory Analysis Sara Vera Data 2.0 April 30, 2013
- 2. In this presentation: 1. Causes Data Overview 2. Quantitative Methods • K-means Clustering • Decision Trees 3. Qualitative Methods • Surveys • Interviews 4. Results
- 3. Causes Data Overview • Basic demographics from Facebook Connect • Gender, Age, Location • Behavioral data from Causes’ website • Actions taken, Campaigns followed • Surveys and Interviews • Third party source to validate our demographic information
- 4. This Research Quantitative Analysis • 200,000 Users on our website in the last 6 months 1. K-Means Cluster Analysis 2. Decision Trees
- 7. Qualitative Analysis 1. Surveys – 1,466 respondents 2. Interviews - 65 participants
- 8. Goals 1. Understanding how people use Causes.com • What inspired you to use Causes? 2. Learning about campaign activity beyond Causes’ website • How active are your friends in protests, petitions, boycotts, etc.? 3. Learning as much as possible about a wide range of topics • First things first: I want to learn more about you! Can you describe your “typical day” (if there is such a thing)? • How big of a role does activism, philanthropy, or volunteering play into your everyday life? Has that changed over time?
- 9. Results: Six User Personas 1. 2. 3. 4. 5. 6.
- 10. Future Directions • Give a face/personality to our users • Engage users with tailored approaches • Identify our ideal user types and find effective ways to engage them
- 11. Thank You
- 12. Who are Causes users? An Exploratory Analysis Sara Vera Data 2.0 April 30, 2013 Hi. My name is Sara Vera and I am a data analyst at Causes, which was founded by Sean Parker and Joe Green as the first Facebook app but recently became a separate platform as Causes.com. We are a website that uses social networking tools for a multifaceted approach to organizing campaigns for social change. The purpose of the research that I’m going to explain today is to understand our users and how to optimize the Causes platform to meet their needs -- since our overarching goal at Causes is to mobilize people to take offline action and get involved in the issues that they care about, whether it’s worldwide, nationwide or in their local community. In this presentation: 1. Causes Data Overview 2. Quantitative Methods • K-means Clustering • Decision Trees 3. Qualitative Methods • Surveys • Interviews 4. Results In this presentation, I am going to explain steps we took to create a picture of our users. I will go over a broad description of the data, our mixed-methods approach – which began with clustering analysis and decision trees and ended with surveys and interviews. I will finish with a description of the resulting personality types of our users. Data Overview 1. Basic demographics from Facebook Connect • Gender, Age, Education 2. Behavioral data from Causes’ website • Actions taken, Campaigns followed 3. Surveys and Interviews 4. Third party source to validate our demographic information Causes collects broad demographic and geographic information through Facebook Connect when users create a Causes account. Most of our quantitative information, however, is collected from Causes’ website. We know which actions a user has taken, which campaigns they are following, how many people they’ve recruited to join a cause and more.
- 13. We supplement our quantitative information with surveys and interviews that cover a wide range of topics from how people use Causes to how people get involved in campaigns or volunteer activities in their local communities. Quantitative Analysis • 200,000 Users on our website from June to December 2012 1. K-Means Cluster Analysis 2. Decision Trees For this research, we subset our data to 200,000 users who used Causes between June and December 2012. From 230 variables, we derived 30 predictive variables by which to start categorizing our users, such as age, income, education, their activity level and topical interests in Causes campaigns. We ran k-means clustering algorithm on our entire data set, which resulted in 6 distinct personality types/personas/clusters. K-means visual: This is a sample clustering with a small sample of our user base (Casual Participants are not included in this sample because they are pretty spread out, forming a cluster in how they are not really represented anywhere else). The bottom two clusters are the Tenacious Veteran Activist and the Self-Assured Millennial, which exhibit drastically different behavior than the top three clusters--that's why they're "farther" away. It's hard to gauge distances like this, but it’s a cool and informative visualization of what our user base looks like.
- 14. We then used these 6 clusters to train the random forest model to find how predictive each variable is in identifying which cluster someone belongs in. Using the random forest model, we found 30 variables that best predicted which cluster a user belonged to. Decision Tree slide: This is an example of one decision tree. (K-means algorithm will tell us whether you’re a Self-Assured Millennial or the Ambitious Activist, random forest will tell us how predictive a variable is in identifying which cluster someone belongs in) Could have done a linear regression as well, but we found random forest to be more accurate. Surveys • 1,466 Survey respondents Goals 1. Understanding use of Causes.com • What inspired you to use Causes? 2. Campaign activity beyond Causes’ website • How active are your friends in protests, petitions, boycotts, etc.? Using the behavioral and demographic trends we saw through our machine learning classification, we dug deeper into the motivations of our users through surveys and interviews. The survey was 20 questions and we received almost 1,500 survey responses. There were more responses in certain clusters, so we took this into consideration for subsequent surveys to get a representative sample. We asked: • What inspired you to start using Causes?
- 15. • What do you aspire to do on Causes? • In the past two months, have you taken action outside of Causes on something you care about (e.g., signed or started a petition, attended a protest, donated to a political campaign, etc.)? • In general, to what extent do you share political and social views with your friends? • How active are your friends in protests, petitions, boycotts, etc.? • What are the top three qualities that make you care about a cause? Interviews • 65 participants Goals • Learning as much as possible about a wide range of topics • First things first: I want to learn more about you! Can you describe your “typical day” (if there is such a thing)? • How big of a role does activism, philanthropy, or volunteering play into your everyday life? Has that changed over time? Users indicated on the survey itself whether they’d be interested in participating in an interview. We conducted 65 interviews of 3 – 5 questions that lasted 30-45 minutes with approximately 10 people from each cluster. We spent several weeks walking through the notes, using design thinking methodology, and highlighting major themes and differences across clusters. Results: Explain the variables on the left … %"of"Total"Causes"Population 39% 14% 13% 11% 10% 8% Age 46 38 25 59 68 35 Female 23% 41% 43% 52% 78% 52% Income 45A55K 55A65K 35A45K 85A100K 20A35K 55A65K College"Educated 42% 55% 73% 61% 34% 62% #"of"Friends"on"Causes 74 294 278 195 359 102 #"of"Invites"sent 157 254 86 104 943 95 #"of"Actions"Taken 10 25 8 19 42 19 Days"Since"Last"Action 16 30 49 4 2 10 #"of"Actions"in"each"category: """Animal"Rights 0.53 0.72 0.98 1.04 2.95 0.23 """Envronment 0.32 0.98 1.02 1.05 0.95 0.85 """Health 0.75 0.12 0.35 0.82 0.23 0.96 """Women's"Rights 0.75 0.82 0.54 0.49 0.32 0.34 Casual" Activist Self7Assured" Millenial Practical" Activist Ambitious" Activist Organized" Retiree Tenacious"" Activist
- 16. “The ambitious activist” is in his mid-forties, and wants to make a bigger impact on the world than he has so far. He’s passionate, talkative, and enthusiastic about sharing his newfound passion with as many people as possible. “The practical activist” is in his late thirties, he has focused ideas on how he can best affect change in the world and he seeks to find a soapbox where he can share his ideas with an audience. Sites like Causes are supplements to his offline activism, not a substitute. He’s frustrated by the number of inefficiencies he sees in the lack of collaboration amongst organizations running different campaigns all focused on a shared goal. “The self-assured millennial” is in his mid-twenties, he is a recent grad with his Bachelor’s degree who is thinking about signing up for graduate school. As with many in their mid-twenties, he is self-confident and believes that he can play an integral part in changing the world around him. It can be difficult for him to pledge allegiance to just one campaign or organization - if he sees injustice, he wants to get involved, no matter the context. “The organized retiree” has recently retired after a successful career and is not ready to slow down. As a practiced organizer, she approaches her work methodically: she researches legislation, educates herself and stays active through leadership roles in local organizations. She is interested in pushing her own definition of what it means to organize effectively, and is curious to see how her own efforts stacks up against others. “The tenacious veteran activist”: Although she is retired and has a couple health problems that keep her mostly homebound, her enthusiasm for change is as prevalent as ever. She’s been involved in campaigns for social change throughout her life; a lack of mobility has driven her to participate now online. Even though taking action on sites like Causes is not necessarily something she enjoys, she does view her involvement as important and meaningful. “The casual participant”: While she would by no means identify herself as an activist, she does recognize the flaws and frustrations in the world around her. She occasionally visits sites like Causes through invites from her friends, but does not feel any loyalty to the campaigns in which she participates. Her busy and successful professional life in addition to her uncertainty about online action attribute to sporadic participation. Future Directions • Ask “What are their needs?” and structure our product roadmap accordingly • Engage users with tailored approaches • Identify our ideal user types and find ways to better engage them