Driving Toward Impact With Data Science #15ntc | Beth’s Blog

Driving Toward Impact With Data Science #15ntc

Data

Photo by Lee Haywood

Note from Beth: I’m attending and presenting a session about Walking As Work with Ritu Sharma.   As luck would have it, one of the sessions that I really wanted to go to is scheduled at the same time as ours.   It is a session about being Data Savvy featuring colleagues from Crisis Text Line, Global Giving, and DataKind.   So, in lieu of a cloning machine, I’m publishing this guest post by By Friederike Schuur, DataKind Volunteer.

From online shopping to online dating, our daily interactions are increasingly going digital and, as a result, generating a huge amount of data about our behavior. Companies like LinkedIn, Netflix and Amazon use this information to better serve their customers in an efficient way and nonprofits are now starting to do the same thing.

Data science is the art of turning the massive amount of data out there into actionable information. If you think of advancing your nonprofit’s mission like driving a car toward a destination, then data science represents all the tools you’d use to navigate the many twists and turns your organization needs to make to get there.

To understand how well you’re serving your constituents and anticipate their future needs, there are three main types of insights that data science can provide to help your organization work even smarter:

  • Descriptive insights help you measure your past and current activity. It’s your dashboard showing information like your mileage and how fast you’re going.

  • Predictive insights help you anticipate future needs or behaviors. It’s the traffic report telling you to expect delays due to the construction ahead.

  • Prescriptive insights help you save time by suggesting next steps to take. Think of it as your GPS, recommending a new route so you avoid that construction altogether.

For a real life example, let’s take a nonprofit that is at the forefront of using data science to advance its mission. One of DataKind’s partner organizations, Crisis Text Line (CTL), is a free, 24/7 text line available nationwide that connects anyone in crisis to crisis counselors. Anyone can text Crisis Text Line from anywhere in the U.S., at anytime, to receive support. Having handled over 5 million text messages to date, Crisis Text Line has a tremendous amount of information to help them maximize their impact. They already have a data scientist on staff, Bob Filbin, focused on using the organization’s data to inform its daily work. They also recently partnered with DataKind and volunteer Noelle Sio of Pivotal on a DataCorps project to do even more with their data.

Let’s see what descriptive, predictive and prescriptive insights came out of it.

Descriptive Insights – measuring satisfaction

Descriptive insights measure past activity and usually involve counting how many times something happened. It seems simple, but sometimes you’re trying to measure something fuzzy and hard to quantify – like gratitude. During their recent project together, DataKind volunteer Noelle helped CTL get a better view of how helpful the texters found their conversations by measuring gratitude in the texts themselves. She used a technique called text analysis to count the number of “thanking” words (thank, thx, thnx, tks) in the messages. This would be impossible for a human to do in any kind of efficient way because of the sheer volume of messages to look through. Thanks to data science, Noelle was able to take a fuzzy concept like “gratitude” and measure it so CTL can better understand its impact on those it serves.

Predictive Insights – better serving constituents by anticipating needs

Most crisis centers and hotlines face the challenge of prioritizing incoming requests. Given the high volume of texters and their wide range of needs, how can Crisis Text Line quickly respond to those requiring urgent interventions? How can they help direct those with ongoing needs to appropriate long-term support? By analyzing Crisis Text Line’s data of past conversations between texters and crisis counselors, Noelle discovered that people who text in more than four times or use words like “school,” “friends,” or “hurt” in their texts are likely to become “repeat texters.” Repeat texters may indicate an individual that needs to be directed to long-term support. By identifying potential repeat texters early on, Crisis Text Line can better triage incoming requests and better respond based on the individual’s needs.

Prescriptive Insights – Providing customized support in a scalable way

Most crisis centers and hotlines face the challenge of prioritizing incoming requests. Given the high volume of texters and their wide range of needs, how can Crisis Text Line quickly respond to those requiring urgent interventions? How can they help direct those with ongoing needs to appropriate long-term support? By analyzing Crisis Text Line’s data of past conversations between texters and crisis counselors, Noelle discovered that people who text in more than four times or use words like “school,” “friends,” or “hurt” in their texts are likely to become “repeat texters.” Repeat texters may indicate an individual that needs to be directed to long-term support. By identifying potential repeat texters early on, Crisis Text Line can better triage incoming requests and better respond based on the individual’s needs.

With further analysis, Crisis Text Line may be able to not only predict who is likely to be a repeat texter, but also to make automated suggestions to counselors on what to do about it. Once Crisis Text Line further investigates the different types of repeat texters and their varying needs, they could then prompt their crisis counselors with suggested actions to take. For example, the crisis counselor might be prompted to ask certain questions or suggest certain resources based on the texter’s needs, thus enabling Crisis Text Line to provide personalized support on a massive scale.

Ready to start YOUR data science journey?

Many organizations worry they’re not ready for “big data” or “data science”.  However, time and time again, we’ve seen that organizations that don’t consider themselves data companies have huge opportunities to use data science in their missions.  Like Crisis Text Line, your organization can use data science to improve its programs, better anticipate needs, or scale your services while still providing customized support to your constituents.

Here’s how to get started:

  • Learn more (and meet Bob from Crisis Text Line!) at DataKind’s panel session at the NTEN Conference March 5th at 10:30am.

  • Ask questions via email, on Twitter or in the comments below!

  • Check out the DataCorps program and apply to work with a team of DataKind volunteers to advance your organization’s mission by using data science.

  • Attend DataKind Chapter events in your area to mix and mingle with fellow data-driven organizations and data-loving volunteers like myself!

No matter how you choose to move forward, the best way to get going on your data science journey is to start now.  Have fun and we’ll see you out there!

Get Started!

 

 

4 Responses

  1. Best of luck to you and Ritu for your session at NTEN! GuideStar is also there boothing.
    Using Crisis Text Line as an example of how big data works really helps me understand it better, I will refer people who are new to big data to this article!
    ~Courtney Cherico, communications coordinator, GuideStar
    Our Blog: trust.guidestar.org

  2. Carl Hardy says:

    thanks for the post good content…………..

  3. I love this guest post and it is a perfect companion post to Deborah Elizabeth Finn’s NTC post and suggestion that the big new job title for the nonprofit sector is “data analyst”: https://deborahelizabethfinn.wordpress.com/2015/03/11/nptech-labor-market-alert-the-big-job-title-of-2015-will-be-data-analyst

    I really appreciate the examples from CTL and love how open they are (here in the guest post and at the 15NTC!) with their data insights and practices. Thanks for featuring them on the blog, Beth!

  4. Beth Kanter says:

    Hi Amy

    I really wanted to attend this session, unfortunately it was scheduled at the same time as my session on walking :(. So next best thing was to get a guest post!

Leave a Reply