Manhattan Condo Sales Prices & Annualized Rental Returns - Snapshot in 2015

Is your real estate agent giving you data?

I compiled the following report on a monthly and quarterly basis between 2014 & 2015, using 3-month trailing data from publicly available closed real estate sales & rental data repository. I subscribed to StreetEasy as a real estate professional which provided data in .csv format. This was only step one.

Data Cleanup (The Stupid Way)

The StreetEasy data download only allowed up to 250 rows per extraction. There were a bit less than 2500 records for each snapshot, so I did 10 extracts and then combined them manually into one single Excel worksheet.

The data was really dirty. It was said that the state paid people minimum wage to enter real estate transaction data into the state registrar and there were tons of data that didn't make sense. Some listed prices seemed to be off by 2 decimal points, addresses were misspelled, ZIP codes that didn't belong to Manhattan, etc. For any listed record that I didn't feel comfortable correcting manually, I would just take it out. There were enough data points so taking a few shouldn't have major impact. This was a very laborious process for the first extract. I would save this "cleaned" extract as a reference file that can be used to clean subsequent extracts. Manhattan real estate properties moved fast but it wasn't so fast that I had to redo the cleaning for most of the closings.

Geocoding

The street address of each record was broken up into "Address", and "ZIP". The City and State were assumed. So I created a concatenated data called "Concat Add" that can actually be paired to a Longitude and a Latitude.

Microsoft's Bing Maps Dev Center (https://www.bingmapsportal.com/application/) provided me with a license key that I can use on a free geocoding website for an automated geocode generation. Google Maps geocoding capability might be faster or more accurate, but I was never able to get the free key to work with the geocoding website. I am very grateful for Bing Maps and their support for individual developers/entrepreneurs.

Data Viz

Once the geocoding was finished, it was a matter of importing the data into Tableau for some quick and beautiful visualization.