The Pretty Charts thread


#101

Pretty Charts


#102

There may be some short disruption to charts today. The paste bin service, snipt.net, that hosts the csv files that my Pretty Charts run off is shutting down at the end of the year. I’m porting them to gist.github.com, which will also do away with my reliance on Google Sheets for formatted versions and for JSONP support.

This also means that henceforth the data will be available in formatted and raw form here (sneak preview, not finished):

gist.github.com/pinsterdev/0f82 … 369184a01d


#103

Pretty Charts (all raw data here)


#104

Pretty Charts (all raw data here)


#105

Thanks for all your work with this. I find them very interesting.


#106

thanks PS


#107

Bit of a typo in previous … it is the September data released today, 17-Nov (RPPI is two months behind from now on because of its dependence on the PPR):


#108

Pretty Charts (all raw data here)


#109

Pretty Charts (all raw data here)


#110

Pretty Charts (all raw data here)


#111

Thanks for fixing the bug. This is really interesting as usual.

The thing is that the apartment data are now displaying a very pronounced rightward skew now. Median Dublin 2-bed apartment is €1800 but mean is €2131. So there are presumably a small number of very high-priced apartments pulling up the mean.

You don’t see this pattern to the same extent in the 2-bed house category. Median of €1700 and mean of €1845.

I wonder if this is down to short-term or corporate lets in the apartment category? These could be put in a category of their own as they don’t interact with the ‘real’ rental market all that much.

Is there away to clean the data to sort for this?


#112

Yes and No - short term / corporate lets are dragging the whole market upwards


#113

You’re right, this deserves a bit of drilling into. Bear with me…

(P.S. One of the things that annoys me is that I can’t tell when a single ad represents an entire apartment block or portion thereof. This probably throws all the numbers off quite a bit. I don’t think there’s any way around that, unless anyone has a bright idea? …)


#114

So here are some breakdowns for the rentals on Daft on 10-Dec-2016 (Dublin only). First, a tabular version of the data already available from the graphs for this date:

Looking at just the 1- and 2-bed apartments, here are numbers broken into price buckets (price >= bucket label). While both distributions show an expected right skew, there is a hump around the €3,000 mark in both cases:

Dublin, number of 1-bed and 2-bed apartment rentals by monthly rent, advertised on daft.ie 10-Dec-2016

I augmented the stats with a weekly/monthly flag for each advertised rental. Looking at the 53 x 1-bedroom apartments with prices of at least €2,800, I see that 36 of them are priced “per week”, implying short term lets. Almost all are in either the IFSC, The Gasworks, or Alexandra Walk (near RCSI). Of the remaining 17 which are priced “per month”, all but two are in either The Gasworks, or The Cedars on Herbert Park Lane. Basically, if you’re paying more than €2,800 for a 1-bed apartment you probably work for Google or in the IFSC, or you might be a diplomat or a rich medical student.

The 2-beds follow a similar pattern. 80% of the 130 renting for at least €2,800 are weekly lets, with the remainder clustered around the IFSC and Grand Canal Square.

As Skippy3 notes, these drag the average apartment rental price up considerably. The “real” average is much closer to the median, basically within the initial humps of the two graphs. And indeed, we can re-run the earlier stats eliminating apartments renting for more than €2,800 to get the following comparison:


#115

Pretty Charts (all raw data here)


#116

Thanks for the detective work PS. Very useful

Am still not really sure how to segment the data, or even if it should be.

Personally I’ll give more weight to the median data in future particularly for apartments.


#117

I think the most sensible approach is to drop anything priced “per week” from the stats, as they skew the numbers quite badly and clearly represent a different market segment. I’ll look into it for the next iteration. I want to simplify the regular posts as well, and not have so much verbiage and quick links.


#118

Pretty Charts (all raw data here)
Latest versions available by clicking the headings below


#119

Pretty Charts (more info and links here)


#120

Pretty Charts (more info and links here)