Transcript KF4 - Ovulation - Brandeis University

RBIF-0103-G1
Probability and Statistics
Prof. M. Partensky
Brandeis University - Spring 2003
Group Project
Using Statistics and Mathematica
to Analyze Body Temperatures
Vaishali Khamamkar
Timothy Foley
Human Body Temperature
as a Predictor of Ovulation
• Tracking body temperature is part of an effective,
low-tech. process that is used to predict ovulation
• Many couples have difficulties in getting pregnant
• There is a large body of online data posted, where
women share their experiences, to help themselves
and each other We thank them, and wish to treat
their personal data with the utmost respect
Technical Data and Measurements
• The posted data is based upon daily measurements
of a women’s body termperature
• A Basal Body Temperature thermometer is used,
accurate to 1/10° F
• The captured data shows a statistical Time Series
Results with Mathematica
• We gathered 20-30 days of BBT data for 12 women
• We plotted that data using:
– ListPlot [ ],
MultipleListPlot [ ]
• We used an Epilog to show
– Target Ovulation Date (Vertical Line)
– Cover Line – Avg. Body Temp. (Horizontal Line)
• 12 Graphs are displayed (Cyan Background)
Data Manipulation
Smoothing of Data
• We wanted to smooth out the data in our results
• A Rolling Mean metric is added, that calculates
each day’s temperature as the average of:
– Previous day’s BBT temperature
– Current day BBT temperature
– Next day’s BBT temperature
• 12 new graphs are shown (light-green background):
– Blank lines shown original plotted time series
– Purple lines show smoothed Rolling Mean time series
Results – BBT and Ovulation
• All women show a rise in BBT values of 0.5°- 2.0° F.
within one day of ovulation
• We also note a BBT plateau, followed by a second
rise in BBT, 3-5 days after ovulation
• More information on this process is available at:
http://www.babymed.com
Concluding Remarks
• We found several interesting areas of study related to body
temperature
• We learned things that we did not know or expect
• Our knowledge of Statistics was broadened by this project
• We found Mathematica to be a powerful and useful tool (once
we got the hang of it!)