WhatsApp Group Data Analysis
We all have assumptions we make about our personal use of social media. Wouldn’t it be great to figure out the truth?
I analyzed my family’s WhatsApp group chat for 2021. These are some of the interesting things I found.
1. To begin with there were total of 3385 messages sent by a total of 37 active participants. This tells me that no matter how much we love each other, we are clearly not chatting every day.
2. Of all the messages sent, only 11% were MMS. Yet when there is a new message notification and it’s a video or image, the general feeling is ‘Not again! This group is wasting my data.’
3. 69% of all chats were sent between 6 a.m. and 6 p.m. I don’t know whether to appreciate the minimal disturbance at night or to feel sorry for anyone (bosses come to mind) who lays claim to those hours. Peak hour? Between 9 and 10 a.m.
4. To expand on the previous point, week days were more active than weekends. Hopefully, this can be attributed to people spending weekends together. The most active time all round? Friday between 9 and 10 a.m. Making plans for the weekend perhaps?
5. Of the top 5 most active days of the year, 2 were weddings, 2 were birthdays, and 1 was just a random discussion that just went on and on.
6. The average number of words per message was 14. This implies some real conversations going on.
7. A total of 335 unique emojis were used. That is, 335 different types of emojis. Makes you wonder how many emojis there are in total. Of course the addition of skin tone to some of the emojis has really broadened the pool.
8. The most commonly used word of the year? Happy. Great to see such positive vibes from people who know way too much about each other.
Generally, it turns out the people I thought were the most active on the group may be outdone by others. And as for me, well my actively is rather low, which is both disappointing and reassuring.
To view the Python code used to condense the data, click here to view it on GitHub. I did create an interactive Excel dashboard to summarize everything (picture below), but I will be working on reproducing it in R just to test my skills.
