A predictive model on marital status based on life events
Til Death Do Us Chart is a predictive model developed using machine learning to predict marital status from life events recorded by the National Longitudinal Survey of Youth 1997.
Exploring the depth of this data set's survey questions influenced the decision to create a predictive model. I ran a single linear model on each survey question and used the ten most predictive variables as the basis for the multiple linear regression model used to predict marital status.
Made with: Python, pandas, NumPy, scikit-learn, JavaScript
The 1997 National Longitudinal Survey of Youth follows over 8,000 participants from their teens through their thirties, asking hundreds of questions and recording how the responses change throughout the study. This predictive analysis explores the changes in relationships, cohabitation, and marital status during this window of their lives.
The Longitudinal Chaos Index (LCI) is a measure of the objective chaos of a subject’s love life. The more changes in your cohabitation/marital status, the higher your LCI. The LCI is based on the survey question: "Respondent's marital status in this month in [1994-2016]," which was recorded every month after subjects' fourteenth birthday. There were six possible responses:
Never Married, Not Cohabitating
Never Married, Cohabitating
Married
Legally Separated
Divorced
Widowed
ABOUT THE MODEL
The LCI begins at zero for each subject and increases by one point with any change to the survey question's response. There are two exceptions:
No change: "Never Married, Cohabitating" to "Married"
+2 points: "Married" to "Divorced"
After calculating the survey subjects' LCIs, I categorized them into four qualitative measurements:
Sad and Alone
Happy Together
It’s About the Journey
Train Wreck
A simple model evaluated each survey question and determined which most affected LCI. A multiple linear regression model predicted LCI based on the top ten most predictive survey questions.
The study confirmed that people are hard to predict. Single variables did not strongly correlate with marital status. However, combining variables increased the accuracy of predictions and the utility of the model.
Continue below to calculate your own LCI!
PREDICT YOUR FUTURE
SAD AND ALONE: < 0.8
Sorry, you'll probably be single. Don't adopt too many cats or be that guy who brings out his reptiles at parties.
HAPPY TOGETHER: 0.8 - 2
Congratulations! You're on the fast track to marriage or cohabitation. However, happiness is not guaranteed.
IT'S ABOUT THE JOURNEY: 2 - 6
Despite some hiccups, you'll likely eventually be married or cohabitating. There's light at the end of the tunnel.
TRAIN WRECK: ≥ 6
Whatever you're doing, you're doing it wrong. A chaotic future is ahead. You might be better off sad and alone.