My roommate had the hypothesis that at the end of the month, second-hand mattresses are cheaper because more people are moving out. I analyzed >6k mattress ads in 12 cities and did not find any effect. The analysis gives insight into the variance of mattress prices in different Dresden city districts.
Introduction
One of my new roommates in Dresden is moving out soon. When getting a second hand mattress for the new room they stated the following hypothesis:
Mattresses that are listed on Kleinanzeigen classified at the end of the month are cheaper.
Why tho? My roommate told me that when she moved out (end of the month) there were a lot of cheap second hand matresses in good quality available.
At this point my statisticians mind was triggered: There for sure are cycles to listed mattrasses with peaks at the 30th and 15th of a month. But some people are very lazy, while some are very early and prepared. In my opinion this should smooth out the curves that the peaks should not be visible anymore and have no statistical effect. So I tested it
At this point I want to mention that this is the second time I test one of those everyday assumptions coming up in discussions with friends. The first was testing if german tv phrases their news in easier language during footballevents for a broader audience watching football. You can find the report HERE.
Data
To answer this question, I investigated all matress ads of the 12 German cities with the highest population. The downloaded advertisements were accessed by searching for “mattress” in the “Mattresses” category. This resulted in the following data set
I downloaded the ads using a 4 second waitingtime between opening a new page to be very gentle to the kleinanzeigen servers. This way the script was almost slower then me manually downloading the pages.
Of course bias is introduced since I myself scraped the pages at the end of the month, but due do the high datasize this bias is mitigated atleast a little.
Descriptive Results
To answer the question I first investigated the pressing question to answer the hypothesis:
Price per Month
As we can see there is no difference in matressprices on Kleinanzeigen. The statistical Intuition was right.
But then: Lets to a small tour and explore if we find other patterns to explain matress prices.
Price per Weekday
Of course weekdays also have no effect aswell, but better check especially when you already have the data haha
What for sure must have an effect are citydistricts. More rich areas should also sell more expensive second hand matresses compared to poorer areas.
Price per District
Fantastic, Sociology and segregation for the Win! (Wait what, cut that) To get a better view of Dresden as a city lets match the districts with their geographical location to sketch a quick and dirty spacial plot.
Price per District (Map)
