What's this all about?
We've recently bought an apartment here in New York, and it got me thinking a bit about the property market. There are lots of websites like http://streeteasy.com/ for finding properties, and there are of course resources like http://www.nytimes.com/pages/realestate/index.html for reading about the market, but you never really get the full picture all at once.
I was interested in seeing what an entire years worth of residential property sales looks like In New York, and so I went in search of a data source. Fortunaltey New York City is actually really good at making data publicly available through the its open data initiative https://data.cityofnewyork.us/. And I specifically found this site http://www1.nyc.gov/site/finance/taxes/property-rolling-sales-data.page which provides data for EVERY SINGLE PROPERTY SALE in New York City for the last 12 months.
I decided to start with just Manhattan and downloaded the Excel. And it looks like this:
Oh no, stupid Excel format - guess I'll have to edit it before loading into Tableau.....
...... nope! Tableau 9 to the rescue:
BOOM! What a cool new feature.
Step one complete (I did add a unique ID field as well). Next up I wanted to be able to map all the addresses. The file includes the address, sometimes with apartment number sometimes without, and zipcode but I'd like to be able to plot each address exactly. This can be done relatively easily. First I had to create a standard format for the addresses:
And then I copied the resulting table of addresses out of Tableau and into a batch geocoder. There area number available on-line, many of them free, but they were taking forever so I tried the one from Texas A&M University. There is a small cost but it is a lot faster and less worrisome than the free versions when doing a large number of addresses. Its also a lot quicker than doing it manually (which is what I used to do......).
The resulting table looks like this, and I will blend this back into the workbook.
Note it includes a field called 'MatchType'. Lets see how it did:
Well not bad overall, but some 300+ building addresses appeared in Brooklyn and Staten Island, which is a bit disappointing when you consider that the zipcode was part of the data set.
Anyhow, I'd say the matching is about 90% correct overall which is pretty decent. This becomes clearer when you colour the buildings by their neighborhood tag.
So what else can we see from the data? Well first up there were some 22,282 sales records in Manhattan in the last 12 months. The next thing I looked at was the sales prices to take a look at the most expensive sales:
Now I know that house prices have been going up, but even in Manhattan $3.3 billion seems a little steep. A bit of googling revealed that 240 1st Ave is in fact the address of Stuyvesant Town, and 3 Peter Cooper Road is Peter Cooper Village. The entire STPCV complex of over 11,000 apartments was bought last year by Fortress Investment Group for $4.7bn.
So its clear I'm not looking at just individual homes here. I investigated further and found that the data set included many types of properties:
So I decided to do two things - exclude any building class that didn't sound like it could be residential, all pure rental buildings and also excluded everything greater than $101m which is the current record for the city or $zero. Another problem with this dataset is that its quite difficult to tell the difference between a whole building sale and an apartment or house sale. So to be safe I excluded all sales that were in condo's or co-ops but did not include any apartment names or numbers. Like so:
And this is how the most expensive sales look after, clearly 157 West 57th Street is doing quite well at attracting its share of billionaires (many of whom of course won't even move in):
But while these sales get all the media attention, I wanted to take a look at how much of the market was controlled by these oligarchy sales, and how much by the rest of the market. Time for some bucketing. The peak for sales prices is between $400k and $600k.
Interestingly there is a dip around $1m. Looking into this in some more detail, we see a big spike at $990 and a big fall at $1m. The reason? A New York City 'mansion tax' of 1% kicks in at $1m:
Looking back up to some bigger buckets, the distribution of number of sales and proportion of $ sold looks like this:
So whilst homes costing over $15m only represent less than 1% of all sales, they account for 14% of the total dollar value paid.
How about the view by neighborhood? Here's a plot of number of units sold versus average sale value by Manhattan neighborhood:
Apartment sales in SoHo averaged over $5m each in the last year. The most sales were on the Upper East Side. The cheapest areas (not labelled above) were Harlem West, Inwood and Morningside Heights.
And how about condos versus co-ops? When we were looking to buy an apartment we learned that co-ops are usually cheaper than an equivalent condo. This is because purchasers of co-ops have to get board approval and that in most cases rules out people who are buying investment properties or buy-to-rent, and so the potential market is smaller. Condo's on the other hand have very few restrictions and so are more popular with non-resident buyers who want to buy purely for investment purposes. Most of the new luxury buildings coming on the market are condo's.
The chart below shows the split sales and average prices by building category. As you can see the majority of sales in Manhattan are split almost evenly between co-ops and condos in elevator buildings. And as suspected, the average condo sales price is significantly higher than the average co-op.
Of course these comparisons don't account for things like square feet or number of bedroom, but they still provide some interesting insights.