Maya Thompson — Educational Data Analyst

When people hear “Educational Data Analyst,” they may picture someone working alone with spreadsheets. How do you see the role?

There is certainly quiet technical work, and I enjoy that part. But the leadership value is not the spreadsheet or the query. It is helping the district ask a question that can actually be answered and then helping people use the answer responsibly. Most requests begin with something broad like, “Can you tell us why these students are not succeeding?” Before I touch the data, we have to define the students, the outcome, the time period, and what decision the team is prepared to make.

I do not own the instructional or program decision, but I can improve its quality. I can show where the evidence is strong, where it is incomplete, and where two departments are using the same word differently. That is a form of leadership because it changes how the organization thinks, not just what appears in a report.

How does technology expand your agency as a leader?

Reproducible queries, governed data models, notebooks, and visualization tools let me spend less time rebuilding the same answer and more time interpreting it with people. If a principal and a program director ask for the same measure, they should not receive two different numbers because two analysts filtered the data differently. The technology helps us make the logic visible and repeatable.

Automation also gives me the ability to notice change sooner. Instead of waiting for a quarterly spreadsheet, we can monitor data quality, participation, or service indicators on an appropriate schedule. But that does not mean every measure needs to be real time. Part of my role is helping leaders decide when faster information will support faster action and when it will simply create more noise.

Tell me about a project that felt like a genuine win.

One of our most useful projects involved chronic absence. The district already had attendance dashboards, but the information mostly described what had happened after students crossed a threshold. School teams wanted an earlier opportunity to respond. We worked with attendance staff, counselors, family liaisons, and principals to identify a small set of signals that could be reviewed weekly without labeling students as destined to become chronically absent.

The first version was too complicated. It had more filters and categories than teams could use during a meeting. We reduced it to students whose recent attendance pattern had changed, the context staff needed to verify the record, and a place to document the next supportive action. Several schools began reaching families earlier, and they found recording errors that would otherwise have persisted. The win was not predicting students. It was shortening the time between a meaningful change and a human conversation.

Was there another success that changed how the district approached data?

We retired a collection of competing “official” enrollment reports. Finance, student services, and assessment each had a spreadsheet with slightly different totals. None was necessarily careless; the reports used different dates, school rules, and inclusion criteria. The disagreement surfaced every time leaders met, and people spent more energy defending their number than discussing the issue.

We created a shared definition page showing the purpose of each measure and established one governed enrollment view for common district use. Specialized reports still exist, but their differences are documented. That reduced reconciliation time and, more importantly, changed the question from “Which number is right?” to “Which definition is appropriate for this decision?” It was a data-governance win disguised as a reporting cleanup.

What is the most persistent hurdle in analytical work?

The pressure for certainty. Leaders often need to act quickly, and a clean number can feel reassuring. But school data is produced by real processes: schedules change, staff enter information under time pressure, program definitions evolve, and some important experiences are not represented in the system at all. A precise calculation can still answer the wrong question.

I try not to respond by becoming the analyst who always says, “We need more data.” Sometimes we can provide a responsible directional answer now. I label what is known, what is assumed, and what would change the interpretation. The goal is useful honesty—not false confidence and not analytical paralysis.

How do you help educators engage with data without feeling judged by it?

I begin with a problem they recognize and ask what they would expect to see. Then we look at a small amount of information together and make the definitions visible. I also ask what the data does not capture. That question matters because it shows that professional knowledge and student context are not inconveniences to be removed from analysis.

Language matters too. If a dashboard says a school is “failing,” people become defensive and the conversation narrows. If it shows a pattern, the population included, and the question the measure can support, the team has room to investigate. Data literacy is partly technical, but it is also about creating conditions in which people can admit uncertainty and learn.

How are artificial intelligence tools affecting your work?

They can be useful for drafting code, summarizing documentation, or generating possibilities to test. They can also produce an answer that looks polished and is completely wrong. We do not place confidential district information into unapproved services, and we do not treat generated analysis as evidence until a qualified person validates the data, method, and result.

The leadership issue is accountability. If an AI tool suggests a query or interpretation, someone still has to understand what it did and stand behind the conclusion. I worry when speed is used to justify skipping lineage, bias review, or human judgment. Used carefully, the tools may increase capacity. Used carelessly, they increase the rate at which mistakes become persuasive.

What advice would you give someone who wants to become an educational data analyst?

Develop technical skill, but do not hide behind it. Learn how enrollment, attendance, assessment, special programs, and school calendars actually work. Practice facilitating a question, writing a plain-language limitation, and telling a leader that the requested analysis may not support the decision they have in mind.

And stay curious when the data conflicts with what staff members tell you. Sometimes the record is wrong. Sometimes the experience is local rather than widespread. Sometimes the measure is missing the thing that matters. Your job is not to make the spreadsheet win the argument. Your job is to help the organization learn what is true enough to act on and what it still needs to understand.

Reflective questions

1. How does an analyst exercise leadership without owning the final decision?
2. What should be established before a district builds a new dashboard?
3. How can qualitative knowledge and quantitative evidence correct one another?
4. When is a directional answer appropriate, and how should uncertainty be communicated?
5. What governance responsibilities belong to analysts rather than only system owners?
6. Which uses of AI could strengthen analytical work, and which require clear limits?