Understanding the gender gap in technology with Girls Who Code

Hello, take a minute and think about your team. Think about them in a stand up wherever you see them most often and ask yourself a question - is your team representative of the general population? For most of you, you had an answer pop into your head right away. Today we are going to talk about a gap that relates to that answer and that is the gender gap.

My name is Heather Weigel. I'm the manager of data and analytics. At Girls Who Code today, we're going to talk about understanding the gender gap in tech.

At Girls Who Code, we understand the gender gap through entry level tech jobs. Every year, there are 23,000 entry level tech jobs, only 26% of those roles are filled by women. What this means is only 5,290 entry level positions every year are filled by women in tech. So the question that we're trying to answer is how do we get from 26% to 50% of those jobs being filled by women?

Our mission at Girls Who Code is working to close the gender gap in tech and we're leading the movement to inspire, educate and equip students who identify as girls and non-binary students with computing skills needed for 21st century opportunities. We're doing this through programming that ranges for students from third grade through high school.

To date, our reach has served over 500,000 students in both in-person and virtual programming. We have 185,000 alumni. 50% of our students come from historically underrepresented groups.

Let's talk about our impact for a second - Girls Who Code alumni are earning computer science degrees at seven times the US national average. Girls Who Code alumni from underrepresented race and ethnicities are earning CS degrees at nine times the US national average. And GWC alumni that we can track in the workforce - more than half are working in tech. We are proud of those numbers.

So let's talk about how we do that. We do that through our programming. Our programming starts as early as third grade. A majority of our students are in our clubs which are in-person programs offered either as after school programs or with community based organizations. We also have clubs for 6th through 12th grade students during the summer when the school year ends. We have our summer immersion programs and those are virtual programs, remote programs offered to the students to fill that gap over the summer. We also offer for students that may want to learn on their own self-paced programming, which is also virtual for our 9th through 12th graders.

We offer mentoring corp, which is a peer program that matches high school students with college age students from our college loops program. I'd also like to call out that we have several new college programs as well including our technical interview prep program this year. Yeah, very exciting.

So how do we tie our work to our mission which is closing the gap in tech? And how do we do that with data? That's what I'm going to dig into a little more today.

There are several questions we want to answer about our mission such as:

• What is the size of the gender gap in tech?
• How do we measure the gap?
• How does our programming contribute to the tech jobs pipeline? Which is something that I'm sure many of you are thinking about when we talk about this issue.
• And are we making progress towards our goal? Are we on track towards closing that gap when we're talking about are we on track towards closing that gap?

Our current analysis shows yes, we are. When we answer this question, we're thinking in terms of the year 2030. And here's how we look at that:

We analyze programmatic alumni data. We look at the past participation of our alumni and their outcomes and the tech job market data. This comes from different government and open source data sets. We to date have taken this data and moved it over to a BI tool. We're using Tableau now to centralize it and make it available internally. And we're continuing to mature how we're both analyzing the data and making it available to our staff.

Some challenges we have run into with this include normalize the data. So we have several different programs and what the data looks like across those programs, across participants and even across alumni with their outcomes, that can be very challenging, deciding what is included as a tech job, what isn't, what roles are included? All of that have been challenges for this analysis.

Data quality and testing is another challenge for us with this. We're currently growing in our testing processes to resolve these issues. And then also making this data accessible to everyone within our organization and accessible externally in the future.

And then how do we keep growing in this analysis? So we repeat this analysis on a yearly basis. Over the next year, we're looking to expand, we have a centralized data warehouse which I'm going to talk about in a minute. We will be loading this data in the future into the centralized data warehouse and expanding on the questions we're reporting on and then also looking towards forecast. So not just answering what we've done in the past, but what do we expect in terms of future trends in the job market and what we predict in terms of outcomes. And that's really exciting.

So we are going to gain a lot of insights as we mature in terms of what we can do with this data and how we're analyzing it and storing it and what we're using.

So what challenges are we facing in this work? We're facing challenges that many of you are probably facing as well. And the first is disparate data sources. We have many different tools and systems across the organization. But this challenge has also created a purpose for us and a new workflow which is to create a centralized source of data and create our data warehouse at Girls Who Code.

One of our goals is to be able to drive strategic decision making and set goals from the reports that are ultimately derived. We use the data warehouse as a data source and then bring that data into Tableau. And we want to be able to make decisions for planning and goal settings with those reports.

Some challenges with this and things to consider are external data sources also have sync jobs between them. And that's something we have to consider and reconcile. And then also the data has to be updated on a daily cadence. So we have a daily run and we have to address any issues with the daily run.

So how are we doing this? We are using Matillion as our ETL tool and we are taking those third party data sources, some of them you may recognize if you are in the nonprofit space. We're using a learning management system HQ. And then we have systems like Campaign Monitor, Classy, Almner. And we are extracting data from them using Matillion and loading them into Redshift.

We also have static data that we are storing in an S3 bucket. And we are loading that into Redshift with Matillion. And on the front end, we are using Tableau for business intelligence. And then for that, we connect to an endpoint in Redshift as a data source. So we do have this centralized source of data.

Another challenge we're facing at Girls Who Code is data compliance. We're working with student data. So compliance is a priority and it's a priority for four reasons:

1. To align with emerging federal and state regulation concerning data privacy and data security.

2. There is a relationship between programmatic growth and the compliance landscape. School districts have expanded their expectations of data privacy and security and we have partnerships with those school districts. So we need to meet those expectations to continue our programmatic growth and we don't want any technical challenges to stand in the way of our program growth.

3. We want to continue to build trust with our users. We want to align with our mission and values.

4. We want to educate our students on how data is being used and to teach them to become advocates of and advocate for their own data and their rights with their own data.

So how do we do this? We are minimizing data. We're minimizing the collection of data in a few ways:

• One is forms tracking throughout the organization. We are collecting data in several different ways, we have lots of stakeholders collecting data in forms on different platforms. And now we have a single source of truth for who is collecting what and how. Starting this year, we are moving to a single platform for the forms collection going into the next year. So we can see everything that's being collected. We know each data point that's being collected.

• We are also continuing in data minimization, deleting at the source level. Anything that is a nice to have is being deleted out of the system. And we are creating processes for review for that.

• For data minimization additionally, we have a core data compliance group who is organizing and prioritizing the compliance initiatives. But we found that implementation is not enough, we have to maintain these initiatives. We have to have people that are responsible for making sure that there is continuous maintenance of what we implement. So the implementation is handled by this data compliance core group. And then most of the initiatives become the responsibility of our new data governance council. And it is a collective effort across the organization.

We are also doing this by controlling access and reviewing all of our access policies, also creating access policies in many places.

And then finally, siloing programmatic data because in our case, some of that programmatic data is owned by the school districts. So normally we're talking about knocking down those silos, but we actually need to spin up new instances. So the data is stored based on ownership just quickly. So you understand what that means for us -

• HQ is our learning management system and we needed to create a second learning management system because of this, the data being owned by the school districts, which meant we needed to create a new instance. Their backend is a MySQL database. And this MySQL database also feeds into the data warehouse.

• So we also needed to create new instance of Redshift in. Yeah, we needed to create a new instance in Redshift and a new S3 bucket as well. So there for the club's data, there is a new instance of everything specifically because of the data ownership.

Another challenge is providing actionable insights across the organization and specific challenges we've had with this are:

• Use case timelines - delivering this quickly, right? We have a lot of use cases that are submitted along with ad hoc requests as we go along and we want to deliver this quickly and effectively to the stakeholders so they can act on these use cases.

• Data quality - data is being used in different ways, metrics are being defined differently across the organization. So how we're working on defining that and unifying around those metrics.

• Balancing compliance and analytical needs is particularly challenging for us right now. On one hand, we want to drive insights from all of our data. On the other hand, we want to minimize the data we're collecting, which is very challenging.

• We're also training staff to use a centralized data platform. We have staff that's very interested in data and very eager to dig in on their own. And now we are asking them to use a centralized platform and that is a bit of a cultural shift. It's an exciting one though, it's a great problem to have.

• We are also working to provide up to date information and actionable data. It's not just presenting the data or a summary but really what can you do with it? If you have an outlier, what is the next step? And that is challenging and over the next year, that's really what we're thinking about - ensuring people know what to do when they have the data, how they can act on it, what it means to them and framing that for them as well.

So what are our solutions for use case timelines?

• We are scaling timelines by training staff to help with the development of dashboards, particularly we are having staff go through Tableau training. And when they have expertise in development data, for example, they are helping on the development of those dashboards.

• For data quality, we are changing our testing processes and instituted the data governance council.

• For actionable data, it's really about understanding the why - how dashboards are used and enabling features to identify the outliers and to drill down on the data.

So how can you help with all of the work that we're doing? There are lots of ways that people can help with this, especially at this audience:

• There are opportunities for employee engagement where you can partner where we form partnerships with different companies. A lot of our programs work with different partners. Some of the programs have employees come in to speak about their roles and their experiences, their job experiences, what they do on a day to day basis.

• We also are looking for investments in the work and resources to continue and grow scale programming.

• But I also want to talk about what you can do as an individual. The partnerships I just spoke about, we have a section on our website where you can present to your company how you can establish a corporate partnership. But you can also either fundraise on your own, we have fundraising partner parties, but most of all, we have clubs and we're looking for facilitators, we're looking for people to help judge our challenges, we're looking for people to help speak.

• We have a newsletter - please sign up there. Please find ways to get involved. We're always looking for people and definitely people with your background and expertise.

If you have any questions, please come find me afterwards. I'd love to speak to anyone about more information on how to get involved. Thank you so much for your time.