Our childcare provider sends us text messages with our daughter’s daily schedule: when she took a nap, when she ate, when her diaper was changed. The messages were always structured the same way, and I realized I could extract the information and create some useful graphs. It is now much easier to monitor my daughter’s patterns as she grows and changes (looking at a graph is way nicer than trying to dig through a bunch of text messages on my phone!)
I call this a no-touch tool because I didn’t ask anyone to change anything they were doing. Too often workers are burdened with extra reporting and data entry duties, and I didn’t want to do that. This app was developed opportunisticaly as an attempt to extract and understand data that already exists.
The project had three steps:
- Extract the data from the text messages into a format I could use.
- This ended up being relatively easy, macs have an sql database with all the messages, and I wrote a little query to pull them.
- Parse the raw text data.
- Here I am fortunate that the messages were so regular. It ended up being pretty trivial to write some regex matches to get everything I needed.
- Create an interactive display
- I wanted to make this data easy to look at and share interactively with my wife, so I made a simple shiny webapp (locally hosted only) to explore what was going on.
All the code is on Github. But the data I’m keeping private.
