Whatsapp Analyzer

Or how I data mined chats with my girlfriend.

I'm always looking for new challenges. This time it was to deduce relevant information from the Whatsapp chat messages that I've had with my girlfriend. ☺

Hereunder are the graphs that the code generates. Let’s see what we can deduce from this!

Date vs Number of messages

This graph shows the date versus the number of messages/chats I've had with my girlfriend. As we can clearly see, we started dating at the end of October 2013, hence the start of the x-axis. There is an outlier (120 chats from me) in March 2014, which indicates that something must have happened that was important. Yes it was, at the end of March I was travelling through China. Furthermore, we didn't chat much from July to August 2014. Something is up here, because we were on holiday together. Last but not least, there is an overall exponential growth in messages since the beginning of our dating.

Does this represent love? I don’t think so.

However, it does show that people in a relationship have an exponential growth over time. The second graph shows one person, me, with the date versus the time I send messages to my girlfriend.

Date vs Time

From this, we can deduce that I’m most actively messaging my girlfriend at late afternoon and in the evening. The average time is 16:46, and the most common day is Wednesday. This corresponds with last year’s study schedule, where I mostly had no lectures on a wednesday afternoon.

These data shows overlappings with other activities, which I briefly explained. However, to truly understand the meaning of these data, it is best to gather more time series data. In example, agenda dates, meetings, overlapping data from the other person, etc… However, it is not to be included at the time of typing as time is scarce.

This project was cool to code and I learned a lot. However my biggest shoutout is to my girlfriend. Without her, this never would have happened. ❤

If you want the code, it is on GitHub, go to https://github.com/Yannvl/Whatsapp-Analyzer. It is coded in Python, using Matplotlib, Numpy and Pyplot. To use, read the readme.

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.