A few days before the new year 2018 I saw this tweet from Deutschlandfunk Nova:

It says that 14 % of the particle matter yearly emitted by traffic is equivalent to the emission produced by fireworks during New Year’s Eve. That was quite shocking for me. Last year I wrote a R-package called senseBox, which is a R API wrapper for the senseBox project. SenseBox is a open-data project for collecting enviromental data. In their own words:

The senseBox is a citizen science project, which was developed in the student lab and research laboratory GI@School at the Institute for Geoinformatics at the University of Münster. This project provides do-it-yourself construction kits for sensor stations with detailed instructions for interested citizens. This enables them to take part in a scientific measuring campaign on environmental data with the help of photonic technology.

They also provide an excellent map where you can observe real-time data from every senseBox. All projects are open-source and you find many code snippets on their GitHub page.

Combining the R API wrapper and the news from Deutschlandfunk Nova brings me to the following image, which shows all available senseBoxes (> 400) with particular matter detectors between 31st December 2017 20 pm and 1st January 2018 4 am. We can differ between two values: PM 2.5 and PM 10. Particulate matter (PM 10 and PM 2.5) comprises the mass of all particles contained in TSP (total suspended particulates) with an aerodynamic diameter of less than 10 µm and 2.5 µm, respectively.

PM 10

PM 2.5

You can clearly see the rise in the y-axis around midnight for both values. The differences between the normal emissions is cleary observable. Note that most of the stations are located in germany (CET) and the timestamps are recorded in UTC on the openSenseMap server.

Furthermore, it was striking that one location always provides a constant value of ~ 2000 µg/m³ for PM 10 and ~1000 µg/m³ for PM 2.5. I contacted the maintainer of this senseBox and we will see if they will find the reason for this problem.

So we once more see that open-data and big data analysis can help to understand our envirnonment better and will hopefully help to make the right conclusions for the future.