How family members make each other happy — An exploration using Scattertext

Kaitlyn Zeichick
6 min readApr 7, 2021

This project was completed as a part of the Metis 12-week data science bootcamp in 2021. This blog post walks through how I used Natural Language Processing to identify how family members make one another happy.


I’ve known my whole life that one of the ways my dad shows me and my sisters that he loves us is by bringing us food. In high school he brought me plates of grapes with cheese and crackers when I was anxiously reciting the parts of the cell in preparation for a biology test. Now he brings me cups of mint tea as I scrape, clean, and model data from the office upstairs. While writing this, he came in the room and asked about making a strawberry pie this afternoon.

The point is, making food is one predictable way for my dad to show love, and it’s a reliable way of making me happy. For this project I was interested in exploring if there are societal trends for how family members make one another happy. To do this I’ll be going over my results from three main questions:


I got my data from a company called Megagon Labs. Megagon Labs is an innovation hub that decided to create an app where users could write down what made them happy that day. They wanted to add a chatbot to their app that could interact with their users. So if a user typed in something like, “I got a raise today!” the chatbot might respond with, “Congratulations!”.

In creating the chatbot, they realized that understanding happy moments is a challenging NLP problem. To solve the problem they collected 100,000 responses on Mechanical Turq where people were asked, “What made you happy within the past 24 hours?” and “What made you happy within the past 3 months?” They then made their data publicly available on Github.

How do parents make their children happy?

Before I jump into ‘how’ mothers and fathers make their children happy, I want to share a couple of very interesting findings.

The first is that mothers and fathers aren’t mentioned in equal amounts. Mothers are mentioned more often than fathers by both their sons and daughters.

On top of that, daughters are more likely to mention their parents as a source of happiness than sons, as you can see in the graph below. Parents make up a larger percentage of ‘happy moments’ in the ‘women’ bar than the ‘men’ bar.

How do mothers make their children happy?

To investigate how mothers make their children happy, I created a scattertext plot that shows what words are associated with mothers versus fathers when people are describing their happy moments.

The first thing that jumps out in the graph is the high association of food-related topics with mothers. Words like “food,” “favorite,” “dinner,” “cake,” and “cooked” are frequently brought up in relation to mothers. This means that when mothers make their children happy, food is often involved.

How do fathers make their children happy?

Unlike mothers, fathers are more associated with gift-giving words like “present,” “car,” “surprise,” and “presented.”

One of the more interesting words was “presented” in the bottom right-hand corner. I looked into these texts a little more and realized that it means being given a gift by the father. The terminology seemed a little off to me, so I looked into it. It turned out the nearly every instance of the term ‘presented’ in reference to fathers was coming from responses from India, which was an interesting finding since most of the responses came from the United States.

How do children make their parents happy?

I decided to approach this by comparing daughters and sons.

How do sons make their parents happy?

Sons were highly associated with sports. For example, one respondent wrote that their happy moment was, “My son hit a home run at his baseball game.”

How do daughters make their parents happy?

Daughters were highly associated with words related to babies. After looking into this further, it looks like there were a lot of responses of people saying they were happy because somebody’s daughter got pregnant.

Daughters were also highly associated with activity-related words like dance and read. So although the activities for sons and daughters are both stereotyped, both of them are activities where children are competing or performing, and might be a source of pride for parents.

Lastly, how do husbands and wives make one another happy?

Once again, women were highly associated with food. People were often happy when their wives made food, bought food, or went out to eat food with them.

Terms related to having children, such as sex and pregnancy, also often came up in relation to wives.

On the other hand, the term husband was often associated with the husband bringing something such as flowers or coffee.

Terms such as dates and anniversaries were also highly associated with the term husband.


As a whole, parents make children happy by giving them things such as food and gifts. Children make parents happy by doing activities. And husbands/wives make each other happy by engaging in relationship-y activities like sex and dates and by giving each other things like flowers and food.


Although these results are a reflection of my data, I’d be very cautious in interpreting them.

As I mentioned earlier, my dad is the one who brings food to me and my sisters to make us happy, rather than my mom. But the results from my exploration showed that moms are more often the ones making their children happy through food. I suspect this is because most of the responses came from the United States, and in the United States women are often the caretakers. In my family, my dad was a stay-at-home-dad. With that perspective, I wonder it ‘mother’ and ‘father’ could be replaced with ‘caretaker’ and ‘non-caretaker’.

This was decidedly a social-science oriented data science project. The results by no means reflects all families, but does give a glimpse into family relationships of common families in the United States.

For more information on this project, including code and presentation slides, check out my Github Repository here.