How to Build One Without Code

Vicki Davis (00:01): Happy Tech Tool Tuesday. This is episode 981. We can build our own working data dashboard and protect private student information.

Announcer (00:16): This is the 10 Minute Teacher Podcast with your host, Vicki Davis.

Vicki Davis (00:21): Today’s show is sponsored by Experience AI, the free AI literacy program from the Raspberry Pi Foundation, co-founded with Google DeepMind. Stay to the end. I’ll share how you can start teaching AI literacy in your classroom with confidence. No experience required. Today, I’m talking with my friend, A.J. Juliani. We first met when we worked on a project called the Flat Classroom Project together. It was in Thomas Friedman’s book, The World is Flat. Remarkable innovator, educator, and also a Wall Street Journal and USA Today best-selling author. He’s an instructor at UPenn’s Graduate School of Education, and A.J. has been leading the charge on helping schools use AI tools for data analysis. He’s back on the show to share how over 150 educators in his national U.S. cohort are building their own AI solutions instead of buying expensive ones.

A.J. Juliani (01:21): Vicki, thanks again for having me on. Love the show and love being a part of it and longtime friends. We’ve kind of all evolved over the years of using AI tools. One of the things that we’ve started to see with newer tools, the one that I use all the time is Claude Code. The ChatGPT version would be Codex. Is it’s taken someone who has a little bit of programming knowledge. I can mess around with HTML and CSS, a little JavaScript here or there, and it’s allowed people to build things that you could only dream of before. And so it has removed barriers that were ever in place for anyone, including educators who have had to pay so much money for custom tools for themselves in the classroom and for districts. And so what I’ve done over the last year here is really help educators who want to use it with a purpose and intentionality of creating things for their classroom and district to show them how to do that. And really, it’s just getting them on, walking through some protocols and giving them some steps to get there.

Vicki Davis (02:23): Walk us through what happened in that cohort when educators are not coders. And what happened when they sat down and built an AI dashboard with Claude Code, Replit or whatever? What surprised you about that?

A.J. Juliani (02:36): The first thing that we did was we talked about what the problem was. The problem was in schools, there’s all these different data sets. There’s conversations in PLCs around data, but there’s never any hub to see everything and to ask questions of the data. Now I’m pulling spreadsheets or I’m pulling some graphs, but I can’t interact with the data. Everybody brought up, this is a very big problem that we have and we’re spending a lot of human capital and time doing that. So we first went and we prompted Claude saying, what are 10 ways that we can use data effectively. Giving our school districts or job title, that type of thing. And so it gave us some different things. Then we took that and we brought that over to Claude Code. If you’ve ever used Claude before, I want to build a dashboard for my district focusing on these X number of things. Help me walk through this process and ask me questions to tailor specifically what this data dashboard should look like. From there, Claude really does the work. It acts as an interviewer, interviewing the educator, getting feedback and insight on them. And that’s the real key that I think people miss along the way. You don’t have to prompt it, everything. It can ask you questions and then take your answers to create it. So after that interview process with Claude, it created what I would say would be like the first version of what this data dashboard would look like. The old way it would do it, you’d have to go into the code and analyze some things and that type of thing. Now you could just have a conversation with Claude Code. People call it vibe coding, but it’s very much a conversation saying, I really like this piece of the dashboard. Can you change this? Can we look at this? And what we then did was say, we want to add that question feature a part of the dashboard. So instead of teams just seeing the data and just seeing the connections, why not build that conversation feeder right into it so PLC teams, district teams, communities can look at district data dashboards and actually have conversations with the data, visualize it in different ways. Vicki, people were mind blown. It was one of those things where the entire cohort wanted to go and work on this with their team right away, dive into it and saw the power of what Claude Code could do for anybody who was using it.

Vicki Davis (04:52): Back up, some districts are huge and they have a lot of data, including a lot of personally identifiable information, that PII, that we have to be so careful about. How did that part work?

A.J. Juliani (05:01): One of the things that we did after we created that MVP, that minimum viable Product, the next step was, hey, how can we make this have the best data privacy? What practices do we need to kind of put into place? And a couple of things really popped up. Number one, the data that we’re using, even if it’s from the district, was not uploaded anywhere. Files are read entirely by the browser and sent once to the AI for analysis. Nothing is saved in the server, database, or disk. Once you close or refresh the page, the data is gone. Is data encrypted when you’re using this type of thing? Well, yes. All communication between your browser and the server use HTTPS, so it encrypts the data. Your file can’t be intercepted while being sent. Who can see the data that you upload? So it’s sent normally through something like OpenAI’s API for analysis. And they don’t use API data for model training, but it does pass through their server. So that would be the one kind of data thing that you’d have to look at as a district. Should you upload real student data? You have to de-identify the data before uploading. Replace student names with labels, remove columns with addresses or phone numbers or identifying data. You’re gonna get meaningful analysis without exposing that sensitive information.

Vicki Davis (06:23): It almost seems too good to be true. Like, did people immediately start making decisions and have epiphanies and just go, oh my goodness?

A.J. Juliani (06:29): Almost immediately. The biggest thing that people had questions about were the privacy concerns. They had to bring that to their IT department, go through the process of approving it and talking to the school board. The privacy of doing it in-house is so much better than giving it to a company that you don’t know what they’re going to do with it and how they’re going to use your data. The flexibility of what you could build, what you could create. The number one aha moment that people had, Vicki, was they realized that they were in control of what this looked like, of how much time they spent on it, of how they build it, and how they could connect different departments already working on this.

Vicki Davis (07:08): Did they get any pushbacks from their boards or their IT? You’re not a coder. You don’t know what you’re doing.

A.J. Juliani (07:13): I think that that pushback happened for a good number of people and they had to have conversations and had to do their homework, right? Everything that I just said about the privacy and that type of stuff, you got to bring the documentation, you got to do your homework, that type of thing. But I will say, when you’re talking with districts right now, and you can save money and not have to go through that whole RFP process and the vendors and everything, the business office perked up. The IT department perked up. There was a lot of meaningful conversations with the school board because schools are strapped for money. They’re strapped for funds. And a lot of times they feel handcuffed by some of these programs and software that they are connected with. Vicki, it’s not just for data dashboards. This is for curriculum housing, right? I think we’ve only hit the tip of the iceberg.

Vicki Davis (07:58): So A.J., what are kind of some of the emotions that when educators in your cohort realize they could build tools that used to cost district thousands, if not tens and hundreds of thousands of dollars? Like what kind of emotions happened with these educators?

A.J. Juliani (08:11): Overwhelm. You know, the reason we did the cohort is the whole idea was AI ready school leaders. This notion that AI is all around us. So how do we get people the knowledge where they can start learning it themselves? There’s a lot of overwhelm around it. But then I think the next step after overwhelm was, this is actually not that complicated. You know, a couple of years ago, people were talking about how we have to have the exact right prompt and these different types of things. With a lot of these tools now, that’s not the case. You need to ask questions. You need to have the AI ask you questions. And I want to show people this because most of the educators that I know that they knew were worried about it for the wrong reasons and thinking it was like when the Internet started and I had to figure out, dial up Internet and they had to be like a hacker to use these different tools. The most important skill right now is discernment. Discernment of when to use it, how to use it, and where it’s applicable. You can’t build that discernment skill if you’re not using the tool. And that’s something I think we really learned in the cohort.

Vicki Davis (09:12): What is the biggest mistake you see schools making right now when it comes to AI and student data?

A.J. Juliani (09:17): We’re outsourcing everything. We’re sending it to this AI company or to this company or we’re gonna use this vendor or this tool. We’ll just worry about it there. I like to really keep things in-house as much as possible where we can create custom tools for our purposes, where we’re not at the beck and call of what the vendor decides and their kind of conversations and how they raise price. Then with the student privacy thing, I think still the biggest concern that I see is that we have kids who are under 18 just using the regular version of ChatGPT, using the regular version of these tools and the privacy is terrible. There’s so many red flags there.

Vicki Davis (09:54): So what is the danger of rushing into AI adoption without a strategy.

A.J. Juliani (09:58): It’s the same problem that we had when people maybe rushed into one-to-one tech initiatives, which is if still, if your school is still built on a compliance-based model, tech only fuels that compliance, whether it was laptops or it’s AI. And a lot of times it becomes digital pacifiers. So you’re rushing to use AI for everything. Do you want a world where AI is helping your teachers plan tasks? The AI is helping students do the tasks. And then the AI is helping your teachers grade the tasks. I mean, now we’ve just removed any human from it possible. And so this idea of discernment, it’s the pause. It’s the conversation of, well, why are we adopting this? Is it just helping us save time or is it helping us with better pedagogy? Is it just helping us save money or is it helping with a better learning experience? Is it just helping us look at data more or is it helping us with making decisions better? Those are the discernment kind of questions that we have to have.

Vicki Davis (11:09): The last time we talked about four essentials, human, social, meaning-centered, and language-based, it kind of sounds like you’re still speaking that language of this human and social and meaning-centered and language-based.

A.J. Juliani (11:21): The first thing is when you’re looking at data, it should tell a human story, right? So it should tell a human story. The numbers should relate to something human. Looking at data by yourself, you’re going to miss something along the way. So we should have conversations in social cohorts. It should also be meaning-based. What’s the point of looking at data? Is it helping us with the 10 different things I talked about, which is, is it helping us with student achievement, with attendance and chronic absenteeism, with equity and subgroup analysis, intervention effectiveness, formative assessments, benchmark data, teacher and classroom data, enrollment capacity planning, school climate data, budget resource allocation? These are the meaning behind looking at data. And then the last piece is language-based and understanding that using AI is becoming a language unto itself. You know, just as you would learn another language, it’s understanding, how do I use this? What do I use it for? What are the conversations that I can have with it?

Vicki Davis (12:25): So what are the questions we need to be asking existing vendors to move forward in a world where we can customize our data dashboards for our schools.

A.J. Juliani (12:34): Right now with the vendors, I think the main thing we can say is, Can you give me a CSV? If you can’t give me a CSV, then it’s not a vendor I should work with because everybody can read a CSV. All the AI tools that we do, we can pull in from the CSV. It is just a spreadsheet and you can use it in any, you know, you can Gemini, ChatGPT, Claude, any of them.

Vicki Davis (12:57): What’s your caution of like, just jumping into the wild west of AI dashboards without some thought?

A.J. Juliani (13:03): The biggest caution is when you’re using real data and not de-identifying it. If you do that, you put yourself in a world of trouble with CIPA, COPPA, all the different types of privacy laws out there. And I think that the big piece here for all of us is we have to go through the proper channels. We have to build policies that support this. We have to get it approved by administration, approved by boards, that type of process. One school, Vicki, just to give you a quick example, they connected their gradebook, their attendance, some of their teacher feedback tools inside of Google Classroom. And what it did was it started flagging when a kid was absent for a couple days or wasn’t turning things in or getting some feedback in some of the classroom right away for the teacher and student to have a conference and a conversation. They said it has completely transformed what the instruction looks like because now students know if they’re missing something, they’re going to have a conversation with their teacher right away. I like to say, Vicki, there’s this big quote that said, you know, I want AI to do my laundry and dishes, not my art and writing. And I think about that. What is the laundry and dishes of our role in education? How can AI help us with those laundry and dishes so we can spend more time doing the things that we want to do with our students and taking things off our plate to free up time?

Vicki Davis (14:28): So, A.J. Juliani, tell us how you can find information on your next cohort.

A.J. Juliani (14:33): You can go to my website where I link to it at ajjuliani.com.

Vicki Davis (14:37): Thank you for coming on the show, A.J.

A.J. Juliani (14:39): Thanks so much for having me, Vicki. And if anybody has any questions, I’d love to come back.

Vicki Davis (14:43): Experience AI, co-founded by the Raspberry Pi Foundation and Google DeepMind, sponsored today’s show. Experience AI is a free AI literacy program downloaded more than a million times worldwide. It provides teachers with ready-to-teach lessons with slides, lesson plans, worksheets, and an excellent free AI glossary that I highly recommend downloading first. Within the resources, there are unplugged activities for you to explore with your learners that don’t require a computer at all. Experience AI supports all teachers, regardless of subject area, and doesn’t require any computer science or background knowledge. Get Experience AI free at coolcatteacher.com forward slash experience AI. That’s coolcatteacher.com forward slash experience AI. I’m recommending Experience AI because these lessons help you teach AI literacy with confidence and teach students how these systems actually work. Thank you, Experience AI, for sponsoring today’s show.

Announcer (15:48): Thank you for tuning in to 10 Minute Teacher Podcast. Join us here every weekday and subscribe to the Classroom Matters newsletter. See you later, educator.