Can my texts tell me if I’m too in love?
Recently, I posted about doing analysis on the text conversations with my ex-boyfriend in an attempt to understand our break-up, see post “Improving My Dating Life One Text Analysis at a Time “.
I’ve received a lot of positive feedback on it and one of the readers recommended a book called The Passion Trap. The book talks about how two people are never perfectly balanced in a relationship. By definition, there is always someone who is more in love with the other. What a successful relationship strives to do is to never let the imbalance go too far. When the imbalance starts, the person who is more in love should take some time to work on himself/herself and build a healthy detachment, while the person who acts more distant should make more efforts to reach out to the other.
I think ultimately what led to my break up was not that I texted too much, but rather I lost too much of myself in the relationship. The balance was off and I was the one who couldn’t see anything but him. Only if there is a pre-warning system alerting me that I should take a breather, then maybe the relationship would still be going on.
Equipped with all the past chat history, I went on to work again… not for my ex, but for a happier future. A good number of analytics has been added since the first iterations, but here are some highlights:
- Am I a text gnat ?
Gnat is an annoying little bug that no matter how much you try, you can never get rid of. A text gnat is someone, like me occasionally, who sends multiple texts to their partners while hearing no responses from them at all. Two indicators are added to measure such a tendency:
a) Spatial-adjusted text ratio: The original text ratio calculation only counts for total number of texts. For example, if I text him 2 messages and he texts me back two, the normal ratio would be a perfect 1. Graph below shows two text scenarios where the normal ratio is 1. However, as you can see from the graph below, the order of the text messages matters. In the left-hand case, I send a message and he replies one back. The ratio is a perfect 1–1. On the right hand side, my one message gets two replies from him. Hence, a spatial-adjusted ratio would show a 1–2. Therefore, a lower spatial-adjusted ratio indicates more text gnat-like behavior.
As one can see below, my spatial-adjusted ratio is slightly lower than my normal ratio, but the same story goes that for every message I received, I sent out two messages in return.
b) Text gnat tendency score: the most straightforward way to measure gnat tendency is to track how many times one repeatedly sends a message without response. More specifically, among total conversations, what portion of the conversation starts out because of one’s second or more attempts at reaching out, ie, when one reaches out for the first time, there is no response. Chart here shows that around 40% of the conversations were from my second or more initiation attempts.
2. How much time do I spend thinking about him ?
In all romantic movies, the main characters would often tell their friends that they find themselves spending more and more time thinking about that boy or that girl. In the modern 21st century where my phone is never 6 feet away from me, I texted my ex whenever I was thinking of him. Assuming that he is thinking of me when sending me a text, I can calculate how many hours a day I think about him at some point during that hour. For example, if I send 10 messages, one per hour is drastically different from sending 10 messages all within 10 minutes. Chart below shows the result. I indeed spent more time during the day thinking about him, on average, 1.5 hours more.
3. The Ultimate Quality indicator: how can I know the quality of my text conversation in an instant ?
As I added more and more metric, it was getting harder to untangle all the details. Hence, the Ultimate Quality indicator is developed to let me know instantly what the quality of my text conversation is. It is a scoring system that adds up the following metrics: (1) Rolling word count ratio: Words sent by partner per message / Words sent by me per message (2) Number of texts ratio: Number of texts sent by partner / Number of texts sent by me (3) Hours in touch ratio: Number of hours partner has sent at least one message / Number of hours I sent at least one message. (4) Multimedia ratio: Number of multimedia partner sends / Number of multimedia I send (5) Gnat percentage ratio: the percentage of conversations started out due to my second or more attempts. This score is subtracted rather than added to the ultimate quality indicator. What strikes me the most about this indicator when applied over the conversation with my ex, is that I felt that my text quality had gone up before the breakup while the indicator would have shown that something ominous might be brewing.
4. What’s next — Should I change who I am ?
This project started as a way to better understand a painful breakup, but I don’t want to just use data to analyze the tragic past, I want it to take me into a brighter future. And honestly, I also just enjoy working on improving data analytics.
I want to be this aloof and mysterious person when it comes to texting (or at least to boost my texting score according to my own algorithm), but I instinctively know that I probably can’t put on this show for a long period of time. I shouldn’t hide who I am to make a relationship work. However, perhaps when I see the score is declining steadily, it just means that I’m spending too much time texting and obsessing over my partner. I should take a bit of time to spend with myself, like writing a few more lines of code or catching up with old friends. Maybe this will not save a relationship that is not meant to be, but at least it might remind me to strive for a better, more balanced relationship.
As always, code is available on git. For those who are curious and want to try it out, you can also submit it through this website and you will receive an automatically-run report.
