Day 2: Fluid Patterns

Part I: The Hook

Crime scene investigators don’t just look for fingerprints, they also analyze blood spatter patterns to reconstruct what happened during a crime. The shape, size, and distribution of blood drops can reveal the height of the victim, the direction of movement, and even the type of weapon used.

Today’s Case Briefing: Detectives have found blood evidence at a scene, but the suspect claims the victim simply fell. Your mission: analyze the spatter patterns to determine if the evidence supports the suspect’s story or tells a different tale.

Part II: The Physical Lab

Materials

  • Simulated blood (red washable tempera paint + water, 50/50 mix), 1 squeeze bottle per table
  • Plastic pipettes/droppers, 2 per student
  • Large sheets of butcher paper or poster board, 3 per team
  • Rulers and protractors, 1 set per student
  • Safety glasses, 1 per student
  • Disposable gloves, 1 pair per student
  • Paper towels and cleaning supplies
  • Aprons or old shirts to cover clothing

Carefully view the below video before conducting the lab: 

Procedure

Step 1: Static Drops, The Height Test

  1. Place a sheet of butcher paper flat on the floor.
  2. Hold the pipette at exactly 12 inches above the paper. Release 5 drops onto the paper, spacing them apart.
  3. Measure the diameter of each spatter. Record your measurements.
  4. Repeat from 24 inches and 36 inches. How does the spatter change as height increases?

Step 2: Moving Drops, The Direction Test

  1. Tape a clean sheet of paper to the table surface.
  2. Fill your pipette with simulated blood. While slowly moving your hand from left to right across the paper, release drops steadily.
  3. Observe the shape of the drops. Which direction do the “tails” point?
  4. The tail of a blood drop always points in the direction of travel. This helps investigators determine movement at a scene.

Step 3: Document and Analyze

  1. Photograph or sketch your spatter patterns from Steps 1 and 2.
  2. For each pattern, record: drop height, drop count, average diameter, and shape description.
  3. Compare your static drops to your moving drops. What differences do you observe?
  4. Using only the evidence, could you determine if a person was standing still or moving? Explain your reasoning.

Step 4: Interactive App

  • Click here to access the app. Start on the “Learn More” tab — read how blood drops form ellipses and why sin(θ) = width ÷ length works. Then switch to the “Calculator” tab: enter the width and length (in mm) from your lab drops, hit Calculate, and watch the formula break down step-by-step while the impact animation shows the drop hitting the surface at that angle. Try at least 3 of your lab measurements. Do your hand calculations match the app? Finally, hit “Practice Challenge” — complete all 5 rounds and aim for 5 stars before moving on.

Step 5: Reflection Questions

  1. How did the spatter pattern change as you increased the drop height? What scientific principle explains this?
  2. Why is the direction of the “tail” on a blood drop important for crime scene investigators?
  3. Do you think blood spatter analysis is reliable enough to be used as evidence in court? Why or why not?

Part III: AI & Digital Literacy

Computer Vision: Can AI See What You See?

Computer vision is a branch of AI that trains machines to interpret images and video. Tesla’s Autopilot uses it to read lane markings at 80 mph. Google’s dermatology AI can identify 288 skin conditions from a phone photo. Pathologists use it to flag cancer cells in biopsies. And forensic scientists are using it to analyze crime scene evidence at a scale no human team could match.

How does it work? AI models are trained on labeled images — sometimes millions of them. Google’s medical AI was trained on over 100,000 retinal scans before it could detect diabetic eye disease. A forensic spatter AI would need thousands of images labeled with height, angle, and impact velocity. The AI doesn’t “understand” spatter — it finds mathematical patterns in pixels that correlate with physical variables.

But here’s the catch:  AI sees patterns differently than humans. A human investigator uses context, experience, and gut instinct built over years. An AI only has pixels — no sense of smell, no understanding of the room, no ability to notice that something “feels off.” This means AI can be incredibly fast and precise, but also make errors that any human would catch instantly.

Keeping the above information in mind, carefully view the below video: 

Activity 

Because you used Teachable Machine in the previous session, this activity should feel familiar. You will build on what you already learned as you apply the tool to a new type of visual data. Your facilitator will guide you through this activity.

Step 1: Revisit Teachable Machine

  • Click here to open Teachable Machine and choose Image Project. Because you already used Teachable Machine in the previous session, this part should feel familiar. This time, you will apply that same tool to a new kind of visual evidence: blood spatter patterns. As you begin, remember that Teachable Machine does not understand the science of blood spatter the way you do. Instead, it learns by finding visual patterns in the images you give it.

Step 2: Build Your Spatter Classes

  • Create three image classes to represent the patterns you studied in the lab: Low Angle Drip for tall, thin drops, Medium Impact for rounder drops, and High Impact Spatter for small, scattered dots. Your goal is to help the AI notice the visual differences between these categories. As you set up your classes, think about how clear and consistent each category really is. If two classes look very similar, the model may struggle to tell them apart.

Step 3: Collect Training Images

  • Using your phone or webcam, take photos of the spatter samples you created during the lab. Try to collect about 15 to 20 images for each class. As you do this, vary the distance, angle, and lighting so your model sees a wider range of examples. This is important because AI systems often perform better when they are trained on data that is more diverse and representative. As you collect images, think about how your choices as the trainer may shape what the model learns.

Step 4: Train and Test the Model

  • Train your model, then test it using a spatter sample that you have not already photographed. Your job is to see whether the AI can correctly classify the pattern based only on pixel information. As you test, record where the model performs well and where it makes mistakes. Pay attention to whether certain categories are easier for the AI to recognize than others and consider what that might reveal about the quality or variety of your training data.

Step 5: Compare Human Analysis to AI Analysis

  • Now compare the AI’s classification process to the method you used earlier in the lab. You classified spatter patterns using width-to-length measurements and observation. The AI classified them using image-based pixel patterns. Think about which method seemed more accurate, which was faster, and what each method may have noticed best. This step is meant to help you reflect on the difference between human reasoning and machine pattern recognition and to consider when AI might be useful and when human interpretation still matters most.

Step 6: Interactive App

  • Reopen the app and go straight to “Practice Challenge.” This time, try to beat your Phase 1 score — 5 random blood drop scenarios, each with different width/length combos. If you get stuck, flip to the “Calculator” tab to check your math, then come back and finish the round. Can you earn all 5 stars without peeking? Once you’ve nailed it, you’re ready to train your AI spatter classifier.

Step 7: Reflection Questions

  1. What is one advantage AI has over human investigators when analyzing images? What is one disadvantage?
  2. If you were training an AI to analyze blood spatter, what kinds of data would you need to collect? How much data do you think would be enough?
  3. Should AI ever make the final decision in a criminal case, or should a human always have the last word?

Previous Lesson
Back to Workshop
Next Lesson