Generating Student Reports with ClassDojo and Formative Assessments

How I generate skill-specific reports from ClassDojo data (and the one trick that saves hours)

Writing student reports is the worst part of my job. Twenty-five reports a term, every report needs to be specific, every parent expects evidence. I used to write them from memory and it took me a full weekend. Now I generate them from ClassDojo data using a workflow that takes one evening, and the centrepiece is what I call the doppelganger class trick. Here is the whole thing.

Step 1: track specific skills, not generic behaviour

Most ClassDojo classes I have seen track generic skills — Participation, On Task, Helping Others. That is fine for behaviour management. It is useless for skill-specific reports. So I create custom skills tied to actual learning objectives: 'Clear Articulation' for speaking, 'Provides Constructive Feedback' for peer review, 'Presents Ideas Clearly' for presentations, 'Uses Evidence' for written work. Whatever I need to report on.

The rule: define each skill clearly enough that I can apply it consistently to every student. Vague skills produce vague reports.

Step 2: award points consistently in real time

Award points during the activity, not at the end of the day from memory. The data integrity collapses if I try to remember who participated well two hours later. ClassDojo's mobile app on a phone or tablet is what makes this practical — I can tap a point on a kid during a discussion without breaking the flow.

Step 3: the doppelganger class trick

This is the part that saves my reports from drowning in noise.

Problem: in my main class, I track everything — Participation, Teamwork, On Task, Off Task, Disruptive, plus the skill-specific stuff. When I export the data, the skill-specific signal is buried in behaviour noise. Sorting through it for every report kills the workflow.

Solution: I create a duplicate 'doppelganger' ClassDojo class — same students, but stripped down to only the skills I want to track for a specific report. 'Speaking Skills' class might have just four skills: Clear Articulation, Effective Communication, Active Listening, Confidence in Speaking. When I run a speaking-focused activity, I open the doppelganger class and award points there. The data is now isolated.

Export from that class, and the spreadsheet contains nothing but speaking-skill data for every student. That is what I write the speaking section of the report from. Same trick for writing skills, collaboration skills, whatever I need.

Step 4: export to Excel, calculate, visualise

Export the doppelganger class to CSV or Excel. Clean up the columns (remove anything I do not need). Calculate average per skill per student. Build a bar graph for each student showing their scores across the four skills. The visualisation is what parents respond to in conferences — they see the bars before they read the prose.

Step 5: use AI to draft the prose

This is where it gets fast. I paste each student's data into ChatGPT (or any LLM) with a template prompt: 'Based on the following scores, write a 100-word paragraph summarising Mia's speaking skills this term. Highlight strengths, name one area for growth, suggest one concrete next step.' I get back a draft. I edit for accuracy and voice. Two minutes per student instead of fifteen.

Privacy caveat: I never paste student names or anything identifying into a public AI. I paste 'Student A' and the scores, get the paragraph back, paste the real name in afterward. This is non-negotiable. Your country's privacy laws may be stricter than you think.

What changed when I moved off this workflow

The whole doppelganger-export-paste-edit pipeline is a workaround for the fact that ClassDojo, Excel, and ChatGPT do not talk to each other. Last term I moved to KiwiBee, and the same workflow runs in one place. ClassSpark lets me track skill-specific behaviour the same way (no need for a separate doppelganger class — I tag points with the skill). Reports Lab generates the AI-drafted report automatically with all the data already attached and no privacy paste step. And the report lands in the parent portal without me copy-pasting into an email. The reports take me an evening instead of a weekend, and I did not have to give up the granularity.

Where this fits in KiwiBee

If end-of-term reports are eating your evenings, KiwiBee adds Reports Lab — formative summaries on autopilot, a connected skills gradebook, and automatic parent-portal delivery. Try it free at app.kiwibee.io.