As noted in Monday’s post, the Property Services Regulatory Authority (PSRA) has launched the Residential Property Price Register (RPPR).  It provides sale price for individual properties for all residential sales, including cash sales.  The data includes the address of each property sold, the date of sale, and the price.  AIRO has been examining the data and has produced a set of interactive data visualizations that summarize the entire dataset by local authority and individual property.  These provide average sales price per local authority and also the median sales price to reveal/compensate for the skewing effect of a small number of very expensive properties, prices of individual properties, box plots of sales per local authority, number of new/secondhand properties.

The data visualisations use the full dataset as published.  Whilst we are confident that the vast majority of the data are robust, it is also clear that there are a small number of errors/issues in the dataset that are having a skewing effect.  These are of two types.  First, some properties have the wrong sale prices attached to them (some checking online reveals that the asking price was radically less than the recorded sales price).  We think that a number of them are probably decimal point errors (e.g. the sale price should be recorded €241k not €24.1m).  Second, several properties have been clumped together and not disaggregated (e.g. an entry of 30 apts for €4.5m, rather than 30 entries for €150K each).  We have not cleaned these errors/issues as we do not know what the correct values should be (the use of the median prices compensates for their skewing effect). We did also try and map all of the properties.   This has proved difficult because the PSRA have not used Geodirectory to clean, standardise and geocode their address format, with a variety of formatting used and many spelling mistakes that make matching to Geodirectory not straightforward.  Even for Dublin we only got a 60% address match on a first pass run, meaning the other 40% would require additional work to match.  Despite these issues, the data does give us a good picture of number of sales and sales prices post January 2011 in different parts of the country.

Eoghan McCarthy and Rob Kitchin

Advertisements