Smartphone Urinalysis Time-Series for Early Signals

ISEF Category: Translational Medical Science

Subcategory: Disease Detection and Diagnosis  ·  Difficulty: Advanced  ·  Setup: Home Setup  ·  Time: Full Year

The Hook

Your urine can change before you feel sick. That makes it a useful signal, but only if you track it the same way every day. A phone and a strip can turn a messy set of color changes into data. Your project asks whether personal trends reveal patterns sooner than one-off tests.

What Is It?

This project uses urine test strips as a simple sensor. Each pad on the strip changes color when it reacts with a substance in urine, like glucose, ketones, protein, leukocytes, nitrite, or pH. Your phone acts like a camera and a rough color meter. You compare each strip to the person’s own past results, not just to a generic chart.

Think of it like watching a stock chart instead of a single price. One reading can mislead you. A time series, which means a set of measurements taken in order over time, can show a pattern. Change-point detection is a way to spot when the pattern shifts more than normal day-to-day noise. In this project, you are not diagnosing disease. You are testing whether daily strip data can reveal meaningful changes that deserve attention.

Why This Is a Good Topic

This is a strong science fair topic because it turns a common low-cost test into a data problem. You can study repeatability, personal baseline shifts, and signal detection without needing a hospital lab. The project connects to real health screening, since urinary markers can relate to infection, hydration, metabolism, and pregnancy complications. You can learn imaging, calibration, statistics, and how to control noise in a real-world sensor.

Research Questions

• How does time of day affect smartphone-measured color intensity on urinalysis strips for the same person? ?
• What is the effect of meal timing on urine glucose, ketone, and specific gravity readings across a 30-day series? ?
• Does a personalized baseline improve change-point detection compared with a fixed manufacturer color chart? ?
• To what extent does phone model or camera app change the measured strip color values? ?
• Which strip pad shows the highest day-to-day variability under home testing conditions? ?
• How does hydration status relate to shifts in urine color and specific gravity over repeated samples? ?

Basic Materials

• Urinalysis strips with pads relevant to glucose, ketones, protein, leukocytes, nitrite, pH, and specific gravity.
• Smartphone with a decent camera.
• White paper or matte white card for a photo background.
• Consistent light source, such as a desk lamp or ring light.
• Ruler or small reference object for image scaling.
• Spreadsheet software for logging results.
• Gloves and disposable cups or clean collection containers.
• Digital kitchen scale or water bottle to help track hydration.
• Notebook for timestamps, meal notes, and symptoms.
• Tape or a fixed phone stand to keep camera distance steady.

Advanced Materials

• Clinical-grade urine analyzer strips from a trusted supplier.
• Color calibration target or color reference card.
• Light box or controlled imaging enclosure.
• DSLR or mirrorless camera with manual settings.
• Spectrophotometer or smartphone spectrometer attachment.
• Image analysis software for color correction and region-of-interest measurement.
• Statistical software for time-series analysis and change-point detection.
• Reference urine simulants for method validation.
• pH meter for cross-checking strip readings.
• Laboratory notebook and coded sample labels.

Software & Tools

• ImageJ: Measures pad color values from images after you standardize lighting and cropping.
• Python: Helps you clean the data, build baselines, and run change-point detection.
• Google Sheets: Lets you log daily readings, symptoms, meals, and timestamps in one place.
• R: Supports time-series plots, outlier checks, and basic statistical tests.
• PubMed: Helps you find review articles and validation studies on urinalysis and urine biomarkers.

Experiment Steps

1. Define the one health signal you want to study first, such as hydration, glucose-related variation, or infection-related markers.
2. Set up a fixed image workflow so every strip is photographed from the same distance, angle, and lighting.
3. Choose the color features you will measure, then decide how you will convert images into numeric values.
4. Build a personal baseline from repeated readings before you test for shifts over time.
5. Plan controls that separate true changes from noise caused by camera settings, room light, or strip lot differences.
6. Select a change-point method that flags unusual shifts, then decide how you will judge false alarms and missed signals.

Common Pitfalls

• Changing lighting between readings, which shifts the strip colors and breaks your calibration.
• Reading the strip at slightly different times, which makes the chemistry look inconsistent.
• Mixing up hydration changes with disease signals, which can create false conclusions from normal day-to-day variation.
• Using the manufacturer chart as the only reference, which hides the benefit of a personal baseline.
• Ignoring camera auto-white-balance and exposure, which causes the same sample to look different across sessions.

What Makes This Competitive

A stronger project goes beyond simple before-and-after photos. You would build a repeatable image pipeline, correct for lighting, and compare personal baseline methods against a fixed chart. You could also test more than one change-point algorithm and report false-positive rates, not just detection results. That kind of careful analysis makes the project look like real sensor validation, not just data collection.

Project Variations

• Track only hydration-related markers, such as specific gravity and urine color, to study how daily routines shift a personal baseline.
• Compare morning and evening urinalysis patterns across different meal types to see whether diet changes the measured signal.
• Test whether a simple phone camera setup or a color-calibrated imaging box gives more stable readings over repeated days.

Learn More

• PubMed: Search review articles on urinalysis, urine biomarkers, and point-of-care diagnostics.
• NIH MedlinePlus: Read patient-friendly pages on urinary tract infection and diabetes screening basics.
• CDC Diabetes: Review public health information on diabetes symptoms, screening, and blood glucose context.
• NOAA Color Science Basics: Use educational material on how cameras, lighting, and color perception affect measurement.
• MIT OpenCourseWare: Search for introductory statistics and signal processing lectures that help with time-series analysis.